Mark Zuckerberg announced Meta Compute, a bet that winning in AI means winning with infrastructure; this, however, means retreating from Reality Labs.
The deal to put Gemini at the heart of Siri is official, and it makes sense for both sides; then Google runs its classic playbook with Universal Commerce Protocol.
Dear Apple, I was, given my interest in virtual and augmented reality, already primed to have a high degree of interest in the Vision Pro, but even so, I appreciate how you have gone out of your way to make sure I’m intrigued. You let me try the Vision Pro the day it was announced , and while I purchased my own the day it shipped (and had it flown over to Taiwan), you recently sent me a demo version of the M5 Vision Pro (it’s definitely snappier, although I don’t like the Dual Knit Band at all; the Solo Knit Band continues to fit my head best). However, the reason I truly know you are trying to win my heart is that not only did you finally show a live sporting event in the Vision Pro, and not only was it an NBA basketball game, but the game actually featured my Milwaukee Bucks! Sure, I had to jump through VPN hoops to watch the broadcast, which was only available in the Lakers home market, but who am I to complain about watching Giannis Antetokounmpo seal the game with a block and a steal on LeBron James in my M5 Vision Pro? And yet, complain I shall: you have — like almost every video you have produced for the Vision Pro — once again shown that you fundamentally do not understand the device you are selling. I’m incredibly disappointed, and cannot in good faith recommend any model of the Vision Pro to basketball fans (or anyone else for that matter). Apple, you are one of the grandfather’s of the tech industry at this point; it’s hard to believe that you are turning 50 this year! Still, you are much younger than TV generally, and sports on TV specifically. The first U.S. television broadcast of a sporting event was a Columbia-Princeton baseball game on May 17, 1939 on NBC; there was one camera accompanying the radio announcer. Three months later NBC televised the first Major League Baseball game between the Brooklyn Dodgers and Cincinnati Reds; this time they used two cameras. All televised sports face a fundamental limitation when it comes to the fan experience: the viewer is experiencing something that is happening in real life 3D on a 2D screen; the solution NBC discovered from the very beginning was to not try and recreate the in-person experience, but to instead create something uniquely suited to this new medium. Two cameras became three, then four, then 147 — that’s how many cameras Fox used for last year’s Super Bowl broadcast . Of course many of those cameras were specialized: included in that number were 27 super slow motion cameras, 23 high resolution cameras, 16 robotic cameras, 10 wireless cameras, and two SkyCams. The job of stitching all of those cameras together into one coherent broadcast falls on the production team, housed in a specially equipped truck outside the stadium; that team coordinates with the broadcast booth to provide a seamless experience where every jump feels natural and pre-meditated, even though it’s happening in real time. It’s a great experience! And, of course, there is the pre-game, half-time, and post-game shows, which used an additional 64 cameras, including 12 wireless cameras, eight robotic cameras, seven augmented reality cameras, and a FlyCam. No broadcast is complete without something to fill the time when the game isn’t on. After all, as advanced as TV broadcasts may be, they still face the fundamental limitation that confronted NBC: how do you translate an in-person experience into something that is compelling for people on their couch looking at a 2D screen? When I first tried the Vision Pro the demo included a clip from an NBA game that was later cut from the demo that shipped with the device (which was the one available in Apple Stores); it jumped out at me at the time : What was much more compelling were a series of immersive video experiences that Apple did not show in the keynote. The most striking to me were, unsurprisingly, sports. There was one clip of an NBA basketball game that was incredibly realistic: the game clip was shot from the baseline, and as someone who has had the good fortune to sit courtside, it felt exactly the same, and, it must be said, much more immersive than similar experiences on the Quest. It turns out that one reason for the immersion is that Apple actually created its own cameras to capture the game using its new Apple Immersive Video Format. The company was fairly mum about how it planned to make those cameras and its format more widely available, but I am completely serious when I say that I would pay the NBA thousands of dollars to get a season pass to watch games captured in this way. Yes, that’s a crazy statement to make, but courtside seats cost that much or more, and that 10-second clip was shockingly close to the real thing. What is fascinating is that such a season pass should, in my estimation, look very different from a traditional TV broadcast, what with its multiple camera angles, announcers, scoreboard slug, etc. I wouldn’t want any of that: if I want to see the score, I can simply look up at the scoreboard as if I’m in the stadium; the sounds are provided by the crowd and PA announcer. To put it another way, the Apple Immersive Video Format, to a far greater extent than I thought possible, truly makes you feel like you are in a different place. The first thing that has been frustrating about the Vision Pro has been the overall absence of content; Apple, you produced a number of shows for launch, and then added nothing for months. The pace has picked up a bit, but that has revealed a second frustration: I think that your production stinks! One of the first pieces of sports content that you released was an MLS Season in Review immersive video in March 2024; I wrote in an Update : I have a lot to say about this video and, by extension, the Vision Pro specifically, and Apple generally. Let me work my way up, starting with the video: it’s terrible. The problem — one that was immediately apparent before I got into all of the pedantry below — is that while the format is immersive, the video is not immersive at all. This is the big problem: This is a screenshot of a stopwatch Mac app I downloaded because it supported keyboard shortcuts (and could thus use it while watching the immersive video). There are, in a five minute video, 54 distinct shots; that’s an average of one cut every six seconds! Moreover, there wasn’t that much gameplay: only 2 minutes and 32 seconds. Worse, some of the cuts happen in the same highlight — there was one play where there was a sideline view of the ball being passed up the field, and then it switched to a behind-the-goal view for the goal. I actually missed the goal the first time because I was so discombobulated that it took me a few seconds to even figure out where the ball was. In short, this video was created by a team that had zero understanding of the Vision Pro or why sports fans might be so excited about it. I never got the opportunity to feel like I was at one of these games, because the moment I started to feel the atmosphere or some amount of immersion there was another cut (and frankly, the cuts were so fast that I rarely if ever felt anything). This edit may have been perfect for a traditional 2D-video posted on YouTube; the entire point of immersive video on the Vision Pro, though, is that it is an entirely new kind of experience that requires an entirely new approach. I had the exact same response when you released a video of a Metallica concert last March : As for the concert itself, the video was indeed very cool. The opening shot following James Hetfield walking into the stadium was very compelling, and, well, it was immersive. And then you cut to another camera angle, and while that camera angle was also immersive, the video as a whole no longer was. What followed was a very enjoyable 30 minutes or so — I’ll probably watch it again — but it felt like a particularly neat documentary, not like I was at a concert. You had a monologue from each member of the band, you had shots of the crowd, you had three songs, all, as Apple proudly noted in their press release, shot with “14 Apple Immersive Video cameras using a mix of stabilized cameras, cable-suspended cameras, and remote-controlled camera dolly systems that moved around the stage.” That means the final product was edited together from those 14 cameras and the four interviews, which is to say it was a produced artifact of a live experience; at no point did I feel like I was at the concert. News flash: I didn’t watch the video again. I’m just not that interested in a TV-style documentary of Metallica. I added: We are nearly two years on from that introduction, and over a year beyond the actual launch of the Vision Pro, and there has yet to be an experience like I envisioned and thought was coming. What is frustrating is that the limiting factor is Apple itself: the company had 14 Apple Immersive Video cameras at this concert, but what I want is only one. I want an Apple Immersive Video camera planted in the audience, and the opportunity to experience the concert as if I were there, without an Apple editor deciding what I get to see and when. Needless to say, you probably already know why I thought Friday’s telecast was a big disappointment. I understand, Apple, why it’s not easy to record or even take a screenshot of a copyrighted game; please bear with me while I describe the experience using text. When I started the broadcast I had, surprise surprise, a studio show, specially tailored for the Apple Vision Pro. In other words, there was a dedicated camera, a dedicated presenter, a dedicated graphics team, etc. There was even a dedicated announcing team! This all sounds expensive and special, and I think it was a total waste. Here’s the thing that you don’t seem to get, Apple: the entire reason why the Vision Pro is compelling is because it is not a 2D screen in my living room; it’s an immersive experience I wear on my head. That means that all of the lessons of TV sports production are immaterial. In fact, it’s worse than that: insisting on all of the trappings of a traditional sports broadcast has two big problems: first, because it is costly, it means that less content is available than might be otherwise. And second, it makes the experience significantly worse . Jump ahead to game action. The best camera was this one on the scorer’s table: I have, as I noted, had the good fortune of sitting courtside at an NBA game, and this very much captured the experience. The biggest sensation you get by being close to the players is just how tall and fast and powerful they are, and you got that sensation with the Vision Pro; it was amazing. The problem, however, is that you would be sitting there watching Giannis or LeBron or Luka glide down the court, and suddenly you would be ripped out of the experience because the entirely unnecessary producer decided you should be looking through one of these baseline cameras under the hoop: These are also not bad seats! I’ve had the good fortune of sitting under the basket as well. These are the seats where you really get a sense of not just the power but also the physicality of an NBA game: I would gladly watch an entire game from here. But alas, I was only granted a few seconds, before the camera changed again. This was absolutely maddening — so maddening, that I am devoting a front page Article to a device no one but me cares about, in the desperate attempt to get someone at your company to listen. What makes the Vision Pro unique is the sense of presence: you really feel like you are wherever the Vision Pro takes you. In other words, when I’m wearing the Vision Pro, and the camera actually stays fixed — like, for example, when you set up a special fourth camera specifically for the Lakers Girls performance, which I think was the single longest continual shot in the entire broadcast — I get the sensation of sitting courtside at Crypto.com Arena, and it’s amazing. Suddenly $3,499 feels cheap! However, when I’m getting yanked around from camera to camera, the experience is flat out worse than just watching on TV. Just think about it: would it be enjoyable to be teleported from sideline to baseline to baseline and back again, completely at the whim of some producer, and often in the middle of the play, such that you have to get your bearings to even figure out what is going on? It would be physically uncomfortable — and that’s exactly what it was in the Vision Pro. What is so frustrating is that the right approach is so obvious that I wrote about it the day you announced this device: one camera, with no production. Just let me sit courtside and watch an NBA game. I don’t need a scoreboard, I can look up and see it. I don’t need a pre-game or post-game show, I can simply watch the players warm-up. I don’t need announcers, I’d rather listen to the crowd and the players on the court. You have made a device that, for this specific use case, is better than TV in every way, yet you insist on producing content for it like it is TV! Just stop! There will be more games this year; from your press release last October : Basketball fans will soon be able to experience NBA games like never before in Apple Immersive on Apple Vision Pro, with a selection of live Los Angeles Lakers matchups during the 2025-26 season, courtesy of Spectrum SportsNet. Viewers will feel the intensity of each game as if they were courtside, with perspectives impossible to capture in traditional broadcasts. The schedule of games will be revealed later this fall, with the first game streaming by early next year, available through the forthcoming Spectrum SportsNet app for Vision Pro. That schedule was announced last week , and there are six games total (including last Friday’s). Six! That’s it. I get it, though: producing these games is expensive: you need a dedicated studio host, a dedicated broadcast crew, multiple cameras, a dedicated production crew, and that costs money. Except you don’t need those things at all . All that you need to do, to not just create a good-enough experience but a superior experience, is simply set up the cameras and let me get from the Vision Pro what I can’t get from anything else: the feeling that I am actually there. And, I would add, you shouldn’t stop with the Lakers: there should be Vision Pro cameras at every NBA game, at every NFL game, at every NHL game, at every MLB game — they should be standard issue at every stadium in the world. There should be Vision Pro cameras at every concert hall and convention center. None of these cameras need a dedicated host or announcers or production crew, because the Vision Pro isn’t TV; it’s actual presence, and presence is all you need. $3,499 is a lot of money for physically uncomfortable TV; it’s an absolute bargain if it’s a way to experience any live experience in the world on demand. But, alas, you refuse. So nope, I still can’t recommend the Vision Pro, not because it’s heavy or expensive or has an external battery, but because you, Apple, have no idea what makes it special.
Welcome back to This Week in Stratechery! As a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle; highlighted links are free for everyone . Additionally, you have complete control over what we send to you. If you don’t want to receive This Week in Stratechery emails (there is no podcast), please uncheck the box in your delivery settings . On that note, here were a few of our favorites this week. This week’s Stratechery video is on Netflix and the Hollywood End Game . Will AI Replace Humans? Dorm room discussions are not generally Stratechery’s domain, but for the terminally online it’s getting harder to escape the insistence in some corners of the AI world that humans will no longer be economically necessary, and that we should make changes now to address what they insist is a foregone conclusion. I disagree : humans may be doomed, but as long we are around, we’ll want other humans — and will build economies that reflect that fact. More generally, as we face this new paradigm, it’s essential to remember that technology is an amoral force: what is good or bad comes down to decisions we make, and trying to preemptively engineer our way around human nature guarantees a bad outcome. — Ben Thompson The Future of Power Generation. Through the second half of last year it became clear that one of the defining challenges of the modern era in the U.S. will be related to power: we need more than we can generate right now, electrical bills are skyrocketing, and that tension figures to compound as AI becomes more integral to the economy. If these topics interest you (and they should!), I heartily recommend this week’s Stratechery Interview with Jeremie Eliahou Ontiveros and Ajey Pandey , two SemiAnalysis analysts who provide a rundown on what the biggest AI labs are doing to address these challenges, the important of natural gas, and how the market is responding to the most fundamental AI infrastructure challenge of them all. Reminder on that last point: Bubbles have benefits ! — Andrew Sharp What China Thinks of What Happened in Caracas. On this week’s Sharp China, Bill Bishop and I returned from a holiday that was busier than expected and broke down various aspects of China’s response to the upheaval in Venezuela after the CCP’s “all-weather strategic partnership” with the Nicolás Maduro regime was rendered null and void by the United States. Topics include: A likely unchanged calculus on Taiwan, a propaganda gift, questions about oil imports, and Iran looming as an even bigger wild card for PRC fortunes. Related to all this, on Sharp Text this week I wrote about the logic of the Maduro operation on the American side, and the dizzying nature of decoding U.S. foreign policy in a modern era increasingly defined by cold war objectives but without cold war rhetoric. — AS AI and the Human Condition — AI might replace all of the jobs; that’s only a problem if you think that humans will care, but if they care, they will create new jobs. Nvidia and Groq, A Stinkily Brilliant Deal, Why This Deal Makes Sense — Nvidia is licensing Groq’s technology and hiring most of its employees; it’s the most potent application of tech’s don’t-call-it-an-acquisition deal model yet. Nvidia at CES, Vera Rubin and AI-Native Storage Infrastructure, Alpamayo — Nvidia’s CES announcements didn’t have much for consumers, but affects them all the same. An Interview with Jeremie Eliahou Ontiveros and Ajey Pandey About Building Power for AI — An Interview with Jeremie Eliahou Ontiveros and Ajey Pandey about how AI labs and hyperscalers are leveraging demand to build out entirely new electrical infrastructure for AI. Notes from Schrödinger’s Cold War — Why the U.S. captured Nicolás Maduro, and the challenge of decoding U.S. foreign policy in an era defined by cold war objectives, but without cold war rhetoric. CES and Humanity Tahoe Icons High-Five to the Belgrade Hand The Remarkable Computers Built Not to Fail China’s Venezuela Calculations; Japan’s Rare Earth Access; A Reported Pause on Nvidia Purchases; The Meta-Manus Deal Under Review A New Year’s MVP Ballot, The Jokic Injury and a Jaylen Brown Renaissance, The Pistons and Their Possibilities The Economy in the 22nd Century, Amoral Tech and Silicon Valley Micro-Culture, What Nvidia Is Getting From Groq
An Interview with Jeremie Eliahou Ontiveros and Ajey Pandey about how AI labs and hyperscalers are leveraging demand to build out entirely new electrical infrastructure for AI.
Nvidia's CES announcements didn't have much for consumers, but affects them all the same.
Nvidia is licensing Groq's technology and hiring most of its employees; it's the most potent application of tech's don't-call-it-an-acquisition deal model yet.
Listen to this post : Pity the paradox of the content producer in the age of AI. On one hand, AI is one of the greatest gifts ever in terms of topics to cover. The 2025 Stratechery Year in Review was, just like 2024 and 2023 (plus a few bangers in 2022 ) completely dominated by AI; my Sharp Tech co-host Andrew Sharp wrote The Definitive Ranking of Tech Company Takeability , and OpenAI was number one with a bullet: OpenAI may or may not be the most important company of the future. There can be no doubt, however, that we are witnessing one of the most takeable enterprises in the history of the world. From the day it was founded — with a non-profit corporate structure that sought to build AGI and then control it themselves “to ensure artificial general intelligence benefits all of humanity” — this company has divided the audience and invited either passionate support or aggressive eye-rolls. On the other hand, LLMs in particular are quite literally content producers! What’s the point of writing analysis when ChatGPT or Gemini or Claude will deliver analysis on demand, about any topic you want? Is this one of those situations like the early web, where the possibility of reaching everyone seemed like a boon but was actually a ticking time bomb for the viability of the traditional publishing model ? I’m actually pretty optimistic about Stratechery Plus’ fortunes for reasons I laid out in last year’s Content and Community : So, are existing publishers doomed? Well, by-and-large yes, but that’s because they have been doomed for a long time. People using AI instead of Google — or Google using AI to provide answers above links — make the long-term outlook for advertising-based publishers worse, but that’s an acceleration of a demise that has been in motion for a long time. What I think is intriguing, however, is the possibility to go back to the future. Once upon a time publishing made countries; the new opportunity for publishing is to make communities. This is something that AI, particularly as it manifests today, is fundamentally unsuited to: all of that content generated by LLMs is individualized; what you ask, and what the AI answers, is distinct from what I ask, and what answers I receive. This is great for getting things done, but it’s useless for creating common ground… Stratechery, on the other hand, along with a host of other successful publications, has the potential to be a totem pole around which communities can form…There is a need for community, and I think content, whether it be an essay, a podcast, or a video, can be artifacts around which communities can form and sustain themselves, ultimately to the economic benefit of the content creator. There is, admittedly, a lot to figure out in terms of that last piece, but when you remember that content made countries, the potential upside is likely quite large indeed. This might, to be fair, be wishful thinking: maybe I’m doomed, but if I’m doomed, probably everyone else is too, particularly when you think about the very long run. It’s the very long run that a fellow content producer, Dwarkesh Patel, considered alongside Philip Trammell in a widely discussed post over the winter break entitled Capital in the 22nd Century : In his 2013 Capital in the Twenty-first Century , the socialist economist Thomas Piketty argued that, absent strong redistribution, economic inequality tends to increase indefinitely through the generations — at least until shocks, like large wars or prodigal sons, reset the clock. This is because the rich tend to save more than the poor and because they can get higher returns on their investments. As many noted at the time, this is probably an incorrect account of the past. Labor and capital complement each other. Wealthy people can keep accumulating capital, but hammers grow less valuable when there aren’t enough hands to use all of them, and hands grow more valuable when hammers are plentiful. Capital accumulation thus lowers interest rates (aka income per unit of capital) and raises wages (income per unit of labor). This effect has tended to be strong enough that, though inequality may have grown for other reasons, inequality from capital accumulation alone has been self-correcting. But in a world of advanced robotics and AI, this correction mechanism will break. That is, though Piketty was wrong about the past, he will probably be right about the future…If AI is used to lock in a more stable world, or at least one in which ancestors can more fully control the wealth they leave to their descendants (let alone one in which they never die), the clock-resetting shocks could disappear. Assuming the rich do not become unprecedentedly philanthropic, a global and highly progressive tax on capital (or at least capital income) will then indeed be essentially the only way to prevent inequality from growing extreme. Without one, once AI renders capital a true substitute for labor, approximately everything will eventually belong to those who are wealthiest when the transition occurs, or their heirs. Or more precisely, it will belong to the subset of this group who save most and most invest with a view to maximizing long-run returns. There is an aspect to this argument that is of the dorm room discussion variety: even if we are approaching the point where AI can create AI (Claude Opus 4.5 appears to be a major leap forward in coding capability in particular), there is still a lot of work to do in terms of making it possible for AI to break out of its digital box and impact the real world via robotics. Part of the thinking, however, is that once AI can create AI, it can rapidly accelerate the development of robotics as well, until robots are making robots, each generation more capable than the next, until everything humans do today — both in the digital but also the physical world — can be done better by AI. This is the world where capital drives all value, and labor none, in stark contrast to the approximately 33% share of GDP that has traditionally gone to capital, with 66% share of GDP going to labor. After all, you don’t pay robots for marginal labor: you build them once…check that, they build themselves, from materials they harvested, not just here on earth but across the galaxy, and do everything at zero marginal cost, a rate with which no human can compete. I get the logic of Patel and Trammell’s argument, but I — perhaps, once again, over-optimistically — am skeptical about this being a problem, particularly one that needs to be addressed right here right now before the AI takeoff occurs, especially given the acute need for more capital investment at this moment in time. First, the world Patel and Trammell envisions sounds like it would be pretty incredible for everyone. If AI can do everything, then it follows that everyone can have everything, from food and clothing to every service you can imagine (remember, the AI is so good that there are zero jobs for humans, which implies that all of the jobs can be done by robots for everyone). Does it matter if you don’t personally own the robots if every material desire is already met? Second, on the flipside, this world also sounds implausible. It seems odd that AI would acquire such fantastic capabilities and yet still be controlled by humans and governed by property laws as commonly understood in 2025. I find the AI doomsday scenario — where this uber-capable AI is no longer controllable by humans — to be more realistic; on the flipside, if we start moving down this path of abundance, I would expect our collective understanding of property rights to shift considerably. Third, it’s worth noting that we have seen dramatic shifts in labor in human history. Consider both agricultural revolutions: in the pre-Neolithic era zero percent of humans worked in agriculture; fast-forward to 1810, and 81% of the U.S. population worked in agriculture. Then came the second agriculture revolution, such that 200 years later only 1% of the U.S. population works in agriculture. It’s that decline that is particularly interesting to me: humans were replaced by machines, even as food became abundant and dramatically cheaper; no one is measuring their purchases based on how much food cost in 1700, just as they won’t measure their future purchases on the cost of material goods in a pre-robotics world. That’s because humans didn’t just sit on their hands; rather, entirely new kinds of work were created, which were valued dramatically higher. Much of this was in factories, and then, over the last century, there was the rise of office work. All of that could very well be replaced by AI, but the point is that the history of humans is the continual creation of new jobs to be done — jobs that couldn’t have been conceived of before they were obvious, and which pay dramatically more than whatever baseline existed before technological change. Like, if I might be cheeky, professional podcaster! Podcasts didn’t even exist thirty years ago, and yet here is Patel — and me! — accumulating capital simply by speaking into a mic and taking advantage of the Internet’s zero marginal cost of distribution, a concept that itself was unthinkable fifty years ago. It’s possible, of course — and to return to my perhaps self-interested and potentially misplaced optimism above — that robots will be better at podcasting than Patel or I. I’m skeptical, though: my experience — and I’ll only speak for myself here — is that the human element is essential in creating compelling content. Sure, sometimes I say “uhm” or “like” or “sort of”, or I get facts wrong, but that’s a feature, not a bug: what I have to say is by definition unique to me, and that is interesting precisely because I am flesh-and-blood, not a robot. 1 Indeed, another way to frame the optimism I have around my career is that the dynamics are the exact inverse of AI: Right now it is individual humans who are uniquely capable to reach audiences at scale; AI, on the other hand, is about scaling compute to deliver results to individuals. Patel and Trammell were, to be sure, talking about the 22nd century, while this is a depiction of the first quarter of the 21st, but I think the desire for a communal experience will persist, and I think those experiences will continue to be organized around other humans, not machines. More generally, I don’t think that this will be limited to podcasting (if such a concept even exists in one hundred years). Consider the most base example: sex. I have no doubt that there will be human-like robots with which you can have sex; I also have even stronger conviction that the overwhelming preference of humans will be to have sex with other humans. And that, by extension, means that all of the courtship and status games that go into finding a lover will persist, and that that itself will be an entire economy all its own. One will not impress a partner with commodity robot-generated goods, no matter how objectively perfect they might be: true value will come from uniqueness and imperfections that are downstream from a human. In fact, I have great optimism that one potential upside of AI is a renewed appreciation of and investment in beauty. One of the great tragedies of the industrial era — particularly today — is that beauty in our built environment is nowhere to be found. How is it that we built intricate cathedrals hundreds of years ago, and forgettable cookie-cutter crap today? That is, in fact, another labor story: before the industrial revolution labor was abundant and cheap, which meant it was defensible to devote thousands of person-years into intricate buildings; once labor was made more productive, and thus more valuable, it simply wasn’t financially viable to divert so much talent for so much time. Perhaps it follows, then, that the devaluing of labor Patel and Trammell warns about actually frees humans up to once again create beauty? Yes, robots could do it too, but I think humans will value the work of other humans more. Indeed, I think this is coming sooner than you might think: I expect the widespread availability of high quality AI art to actually make human art more desirable and valuable, precisely because of its provenance. 2 It’s also worth noting the relative popularity of human-generated content versus AI-generated content. Sora is down to 59 in the App Store, and I count double-digit human-denominated social apps that rank above it. Yes, I get the argument that this is the worst that AI will ever be, but it also will never be human, which is what humans want most of all. This gets at what I found the most frustrating about Patel and Trammell’s point of view: the core assumption undergirding their argument was also about the human condition; it just happened to be negative. Louis C.K., in an October 2008 appearance on Late Night with Conan O’Brien , delivered one of the most incisive and eternal observations about human nature: “Everything is amazing right now and nobody’s happy.” You’ve almost certainly seen this clip, but if not, it’s worth watching in full; Louis C.K. focuses on three incredible technological innovations and how quickly we took them for granted: smartphones, Internet access on planes, and the act of flying itself. It’s certainly a sentiment I can relate to: just in the last 72 hours I have chafed at slow airplane WiFi, complained about jet lag from having literally traversed the globe, and gotten frustrated at an iPhone bug that is sapping my battery. It’s all so terrible, until I remember I have access to anything everywhere, can be anywhere anytime, and oh yeah, can achieve both simultaneously. If anything, you can make the case that technological innovations, by virtue of conferring their benefits on everybody, has actually had the perverse effect of making everyone feel worse off. When I was a child growing up in small town Wisconsin, I had some sort of vague sense that there were rich people in the world, but from my perspective taking my first airplane flight around the age of ten was a source of great wonder, and even provided a sense of status; after all, many of my friends had never flown at all. That was the comparison set that mattered to me. Social media — or, more accurately, user-generated content feeds, which are increasingly not social at all — has completely changed this dynamic. All I or anyone else needs to do is open Instagram to see beautiful people on private jets or on beaches or at fancy restaurants, living a life that seems dramatically better than one’s dull existence in the suburbs or a cramped apartment; never mind that the means of achieving that insight is a level of technological wealth that would have been incomprehensible to the richest person in the world fifty years ago. To put it another way, what Louis C.K. identified in this clip was the extent to which human happiness is a relative versus absolute phenomenon: what we care about is not how much we have, but how we compare. That, by extension, is what drives the technological paradox I noted above: more capabilities, more broadly distributed, has tremendously enriched the world on an absolute basis; the end result, however, has been the dramatic expansion of our comparison set, making us feel more immiserated than ever. 3 This, writ large, is what Patel and Trammell seem to be worried about: sure, everyone may have all of their material needs met, but that won’t be good enough if the price of that abundance is the knowledge that someone else has more. This might not be rational, but it certainly is human! If you assume that the negative parts of humanity will persist in this world of abundance, however, then you must leave room for the positive parts as well, the ones that I wrote about above. Even if AI does all of the jobs, humans will still want humans, creating an economy for labor precisely because it is labor. You can’t make the case that the potential for jealousy ought to drive authoritarian capital controls while completely dismissing the possibility that the prospect of desirability gives everyone jobs to do, even if we can’t possibly imagine what those jobs might be — beyond podcasting, of course. If you want a specific example, consider the rapturous response to Bill Simmons’ 50 Most Rewatchable Movies of the 21st Century episode, which was delightful precisely because it was pure Bill ↩ This isn’t idle talk: I’m encouraging my daughter to pursue art; granted, I’m also working quite hard to build up a store of capital for her as well! ↩ As an aside, this is why galaxy exploration would be a positive, not a negative: out of sight out of mind, just like it used to be. ↩ If you want a specific example, consider the rapturous response to Bill Simmons’ 50 Most Rewatchable Movies of the 21st Century episode, which was delightful precisely because it was pure Bill ↩ This isn’t idle talk: I’m encouraging my daughter to pursue art; granted, I’m also working quite hard to build up a store of capital for her as well! ↩ As an aside, this is why galaxy exploration would be a positive, not a negative: out of sight out of mind, just like it used to be. ↩
Stratechery is on holiday from December 22, 2025 to January 2, 2026; the next Stratechery Update will be on Monday, January 5. In addition, Sharp Tech , Sharp China , and Dithering will all return the week of January 5. Greatest of All Talk and Asianometry will post through the holidays on a reduced schedule. The full Stratechery posting schedule is here . Merry Christmas and Happy New Year!
While I started Stratechery in 2013 while living in the United States, within months I moved back to Taiwan , which means the vast majority of Stratechery content was written from the other side of the world from the companies I covered most closely. The big event this year is that I moved back : Stratechery, now in its 13th year, is once again a U.S.-based publication. Here are the Taiwan 12 Years in Review: 2024 | 2023 | 2022 | 2021 | 2020 | 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 This year was once again dominated by AI, particularly in the context of big tech companies; another consistent theme, however, was the opposite of progress: concern about the United States’ competitive position in manufacturing in particular. It’s striking, but perhaps not surprising, that these two issues are rising in prominence at the same time. Will AI save America, and can America build what it needs to truly reap AI’s benefits? This year Stratechery published 26 free Articles , 109 subscriber Updates , and 39 Interviews . Below is a summary of most of the Articles, Interviews, and a selection of my favorite Updates of the year. The five most-viewed Articles on Stratechery according to page views: Nearly every Article on Stratechery touched on AI; these Articles took a broader view than just one company. The companies spending the most on AI — and with the most to potentially lose — are the biggest tech companies. Plus, the implication of self-driving cars and the end game in Hollywood. One big tech company received special focus this year: Apple, which stumbled in its initial attempt to build AI. America is leading the way in AI, but is finding itself behind in manufacturing; policy makers have to consider the implications of both. Every week (except July and August) I post a Stratechery Interview — in podcast and transcript form — with public company executives, private company founders, and other analysts. ServiceNow CEO Bill McDermott | Uber CEO Dara Khosrowshahi | Snowflake CEO Sridhar Ramaswamy | Google Cloud CEO Thomas Kurian | Meta CEO Mark Zuckerberg | SAP CEO Christian Klein | Nvidia CEO Jensen Huang | Cloudflare CEO Matthew Prince | YouTube CEO Neil Mohan | Booking CEO Glenn Fogel | Asana Founder Dustin Moskovitz | Unity CEO Matthew Bromberg | Atlassian CEO Mike Cannon-Brookes | Rivian CEO RJ Scaringe Anduril CEO Brian Schimpf | Manna CEO Bobby Healy | Tailscale CEO Avery Pennarun | OpenAI CEO Sam Altman in March and October | Plaid CEO Zach Perret | Cursor CEO Michael Truell | Sierra CEO Bret Taylor | Flighty CEO Ryan Jones Jon Yu on YouTube and semiconductors | Daniel Gross and Nat Friedman on AI | Matthew Ball on gaming | Benedict Evans on AI unknowns | Michael Nathanson on streaming in March and December | Dan Kim and Hassan Khan on CHIPS, and Dan Kim on Intel and Nvidia | Eric Seufert on digital advertising in April and November | Patrick McGee on Apple In China | Ben Bajarin on AI Infrastructure | Gracelin Baskaran on rare earths | Michael Morton on e-commerce Some of my favorite Stratechery Updates from 2025: I am so grateful to the subscribers that make it possible for me to do this as a job. I wish all of you a Merry Christmas and Happy New Year, and I’m looking forward to a great 2026! DeepSeek FAQ — DeepSeek has completely upended people’s expectations for AI and competition with China. What is it, and why does it matter? Google, Nvidia, and OpenAI — OpenAI and Nvidia are both under threat from Google; I like OpenAI’s chances best, but they need an advertising model to beat Google as an Aggregator. The Agentic Web and Original Sin — Microsoft is putting forth compelling proposals for the Open Agentic Web. However, the proposal needs digital payments, which will be key to creating a new content marketplace for AI. U.S. Intel — The U.S. taking an equity stake in Intel is a terrible idea; it also happens to be the least bad idea to make Intel Foundry viable. The Benefits of Bubbles — We are in an AI Bubble: the big question is if this bubble will be worth it for the physical infrastructure and coordinated innovation that result? AI’s Uneven Arrival — o1 / o3 points the way to AGI, which is AI that can complete tasks; it may take longer for most companies to adopt them than you might think: just look at digital advertising. Deep Research and Knowledge Value — Deep Research is an AGI product for certain narrow domains; its ability to find anything on the Internet will make secret knowledge all the more valuable. Checking In on AI and the Big Five — A review of the current state of AI through the lens of the Big Five tech companies. Tech Philosophy and AI Opportunity — Positioning AI contenders — and losers — by their tech philosophy and business potential. Content and Community — The old model for content sprung from geographic communities; the new model for content is to be the organizing principle for virtual communities. Facebook is Dead; Long Live Meta — Meta delivered blowout earnings the same quarter that Mark Zuckerberg doubled down on AI; I don’t think it was a coincidence. See also: Sora, AI Bicycles, and Meta Disruption — Sora is going viral, suggesting there is a big opportunity in unlocking creativity. If that’s true, that’s good for humanity — and bad for Meta. The YouTube Tip of the Google Spear — I’ve come to appreciate Google’s amorphous nature; what makes me bullish is the clarity of YouTube’s AI opportunity. OpenAI’s Windows Play — OpenAI is making a play to be the Windows of AI: the all-encompassing platform that controls both hardware supplier and software developers. Robotaxis and Suburbia — Robotaxis are poised to further close the delta between suburbs and the city; the city (and Uber) might never recover. Netflix and the Hollywood End Game — Netflix is driving the Hollywood end game, likely confident it can increase the value of IP, and fend off YouTube. Apple AI’s Platform Pivot Potential — Apple AI is delayed, and Apple may be trying to do too much; what the company ought to do is empower developers to make AI applications. Apple and the Ghosts of Companies Past — Apple is not doomed, but for the first time in a long time its long-term fortunes are cloudy; the time to make change is now. Apple Retreats — Apple’s WWDC was a retreat from not just last year’s WWDC, but potentially a broader reset for the company. That’s why it was a great presentation. Paradigm Shifts and the Winner’s Curse — When paradigms change, previous winners have the hardest time adjusting; that is why AI might be a challenge for Apple and Amazon. iPhones 17 and the Sugar Water Trap — Apple’s iPhone announcement was impressive, but no one was impressed, because Apple is increasingly peripheral to what is changing the world. AI Promise and Chip Precariousness — The AI industry is more exciting than ever, but the chip situation is very precarious and requires drastic action. American Disruption — A new take on Trump’s tariffs, including using a disruption lens to understand the U.S.’s manufacturing problem, and why a better plan would leverage demand, not kill it. Resiliency and Scale — Decreasing transportation and communications costs increases resiliency in theory, but destroys it in practice. The only way to have resiliency is through less efficiency. January 8 : Meta Changes Moderation Policies, Zuckerberg’s Journey — and Mine, The Audacity of Copying Well February 4 : Apple Earnings, OpenAI Deep Research, The Unbundling of Substantiation February 19 : Encryption and the Uneasy Compromise, Netflix Earnings, The Aggregator’s Compounding Advantage March 3 : Microsoft EOLs Skype, Skype’s Founding, Microsoft’s Skype Charity March 5 : Alexa+, A Brief History of Alexa, Amazon — and Apple’s — Mistake March 19 : Nvidia GTC and ASICs, The Power Constraint, The Pareto Frontier April 15 : Meta v. FTC, The Three Facebook Eras, Video Slop and Market Forces May 14 : Airbnb’s New App, Experiences and Services, Chesky’s Founder Mode May 21 : Google I/O, The Search Funnel, Product Possibilities June 3 : Nike on Amazon; Nike’s Disastrous Pivot; Inevitability, Intentionality, and Amazon June 24 : Talent Wars, NBA Money, AI Money July 14 : Google and Windsurf, Stinky Deals, Chesterton’s Fence and the Silicon Valley Ecosystem July 16 : Cloudflare’s Content Independence Day, Google’s Advantage, Monetizing AI July 28 : TSMC Earnings; A16 and TSMC’s Approach to Backside Power; Intel Earnings, Architecture, and AI July 30 : Figma S-1, The Figma OS, Figma’s AI Potential August 27 : KPop Demon Hunters, Sony’s Risk, The Netflix Aggregator September 9 and November 11 on SpaceX’s foray into terrestrial spectrum September 24 : YouTube Restores Suspended Accounts, Free Speech and Cultural Mores, Platform Power November 3 : Google Earnings, Meta Earnings, The Cost of Reality Labs December 10 : Trump Allows H200 Sales to China, The Sliding Scale, A Good Decision
ChatGPT Image 1.5 launched, and while it seems comparable to Gemini's Nano Banana Pro, the product around it shows OpenAI's advantages. Then, Apple v. Epic rolls on.
Listen to this post: Good morning, Today’s Stratechery Interview is with Rivian founder and CEO RJ Scaringe . Last week Rivian held their Autonomy and AI Day , where the company unveiled its plans for a fully integrated approach to self-driving . Rivian is building everything from its own chips to its own sensors — including video, LiDAR, and radar — and if all goes well, the company will supply a multitude of companies, particularly Volkswagen. In this interview we cover all aspects of Rivian, including the long path to starting the company, production challenges, and why partnerships with Amazon and Volkswagen are so important, and point to relationships in the future. We also dive into autonomy, and why Rivian is taking a different path than Tesla, plus I ask why CarPlay isn’t available on Rivian vehicles, and what that reveals about their nature. As an aside to podcast listeners: due to a mind-boggling mistake by me, the first 20 minutes of this podcast are considerably lower audio quality. I forgot to hit ‘Record’, so the segment that remains is what the Rivian PR represenative captured on her phone. I’m incredible grateful for the save. As a reminder, all Stratechery content, including interviews, is available as a podcast; click the link at the top of this email to add Stratechery to your podcast player. On to the Interview: This interview is lightly edited for content and clarity. RJ Scaringe, welcome to Stratechery. RJS: Happy to be here. We are here to talk about Rivian and your recent Autonomy and AI Day. Before we get to that, however, I want to learn more about you and your background, and how you ended up sitting with me today. You were, as I understand it, into cars at a very early age. RJS: I’ve been around cars as long as I could remember. As a kid I was restoring and working on cars. I spent time in restoration shops helping and slowly learning how to do more than “help”, but actually really help. And then around the age of, I guess 10-ish, decided I wanted to start a car company. Oh, okay. So there was no like, “Oh, I got a computer and started typing out BASIC”, this is straight cars all the way. RJS: Yeah, I knew I wanted to start a car company. And at that point when you’re a kid, you have no idea what it entails, you have no idea what the business is going to be, but I just knew that it was something I wanted to do and I sort of started charting out my future path with that as the end state goal, with that as the context. So I went and worked as a machinist, I ultimately went to school for engineering, I did a degree in mechanical engineering, then I went and did a master’s and a PhD focused on automotive. And then the day after I finished my PhD, I officially started Rivian. So why did you think it was necessary to do that level of education? Not just a bachelor’s or not just gain experience, but to go all the way through to the PhD? RJS: It was actually pretty intentional. I knew that to start a car company, I was intellectually honest with myself that it would take a lot of money and I knew that I didn’t have any money. So that meant for me to do this and be successful, I would need to get other people to invest money into the idea and typically in the tech space, you could start something with not a lot of capital that you can make a very crude version of your first product… Because it’s software not hardware RSJ: Exactly, and I also knew that I didn’t want to go work for 25 or 30 years to accumulate experience that would make me credible. So I was like, “What’s the fastest path to credibility?”, and my thought was it would be a PhD. I said, “If I get a PhD from a top school” — I went to MIT — I thought that would be some earned credibility that would make it more likely that investors would want to get into the company. I didn’t grow up around venture capital, I had no idea around any of these things, that was my hypothesis. And amazingly, it proved to be a key element in Rivian’s journey, because one of our early investors, one of our earliest large investors, I should say, was someone that I was introduced to through MIT and was an alumni of the school and I was connected with them through the provost, and that was ultimately what led to some of the really critical early capital into the business. So is it the end state that the PhD was totally worth it, but the actual academics was completely incidental to this introduction? RJS: Yeah, and I think as is the case I think with all higher education, the biggest takeaway is to learn how to learn, and to learn how to solve complex problems. I think undergrad, you’re learning how to learn. Graduate school, particularly for technical degrees, you identify a problem and you work really hard to solve that problem, and you have broad responsibility and broad scope on doing that activity, and you build confidence and you build skillsets around problem solving. But the problems are going to change in the course of a life or in the course of your career, the things I was working on 20 years ago have nothing to do with Rivian whatsoever. Well, I’m actually kind of curious about that. What were the things that you were working on 20 years ago, and why aren’t they applicable? RJS: Well, in the case of automotive research, in 2005 the work that was funded, which is the kind of work that you do as a PhD student, so you get sponsored by companies or by grants, was to look at making engines, internal combustion engines more efficient, that was primarily the focus. And so I was working on something called a homogeneous charge compression ignition engine, which is a different type of combustion. We’d compression ignite a pre-mixed fuel air mixture, very hard to get- Like diesel? RJS: It’s like diesel efficiency with gasoline-like cleanliness is the idea. Obviously, it’s not a technology that has any runway and makes any sense in the future. (laughing) I’m hearing about if for the first time right now. RJS: Yeah, so it was an interesting project. It was really a study in software controls, because that was the challenge of this project. But I didn’t take a single piece of that and use it in starting Rivian. Now, that’s different, some folks turn their PhD into the foundation of a business and start a business off the back of it, I had the benefit of just being in the automotive lab, I had the benefit of working closely with car companies. Big, large car companies were funding a lot of the work and it further solidified my view that I didn’t want to go work at one of those companies, and I thought the likelihood of me learning the necessary skills is lower working at one of those places than me learning by doing, by just going and starting a company as a 26-year-old PhD graduate. Right. So if you start out with cars as a child and you’re coming all the way up, at what point did you know that the car — you wanted to start a car company, at what point did you know it was going to be electric and not internal combustion? RJS: That was far less clear. I wanted to make cars very efficient and I wanted to design cars that would essentially help define what the future of state would look like. But when you’re 10, you have no idea what that means. When you’re 20 — and at this point, this was early 2000s, it still wasn’t that clear — and so it didn’t really become clear until I started the business. Even before starting the business, one of the concepts that was competing for what Rivian ultimately would become was this idea I had for a pedal-powered car, which at the time I was thinking could be a hybrid-electric, except the hybrid, it was human-plus-electric drive and amazingly, full circle, that happens with e-bikes. E-bike is the most popular electric— Oh right, yeah. RJS: But 25 years ago, 20-plus years ago, that wasn’t clear that e-bikes were going to be an explosive success as they’ve been. And then I created within Rivian, a skunkworks team that’s now spun out into a new company to actually focus on this pedal and pedal-hybrid electric vehicles. So we have a quadricycle we’re doing with Amazon as a first big customer, but the name of this company is Also, and the idea of this spin-out from Rivian is that if you want to electrify the world, you need to electrify vehicles, but you also need to electrify everything else, and so Also is doing everything else. So Tesla started in 2003. Was there any inspiration or connection there, or is it just incidental that it ended up being kind of around the same time? RJS: Yeah, so they started in ’03, I started in ’09. Of course I was aware of Tesla, but Tesla launched their first car, the Roadster , before I even started Rivian and so they launched the Roadster and then they were working on the Model S , which it doesn’t get talked about a lot, but there was a time when the Model S was considering using a series hybrid architecture as well. Ultimately, they went pure EV but that was in like 2008, 2009, and I started the company just as, my view is there’s going to need to be a lot of successful choices, and I’d been on that mission for a while — what I didn’t expect is just the process of raising capital is really hard. So is that actually where they did help you a lot, just because eventually once they got over the hump and it was a successful venture, did that make it easier for you in the long run? RJS: I think so, and I think Tesla was the existence proof that I’d say more than raising capital, what Tesla did is they showed that electric cars could be cool. RJS: And they did that with the Roadster. So they launched this Roadster, they took a Lotus Elise , they re-engineered it, they made it electric. It was super fast, it was really cool, this was way before anybody had thought about electric vehicles as something that could be fun or fast accelerating. Now it’s hard to believe that 20 years ago this was the case, but at the time they really took electric cars from this perspective like golf carts, to like, “Oh, this can be a highly capable performance machine”, and that just shifted mindset, and that was important. So you start Rivian in 2009, I believe the first vehicle comes out in 2021 . That’s a long time period, what was going on for those 12 years? Are these painful memories? RJS: No, no, no, no, they’re useful memories. In the beginning you have no capital, so you can’t realistically make progress on building something like a car unless you have some level of capital. So if you’re spending $1 million a year, you need to spend 5,000 times more than that to ultimately launch a car, something like that, maybe more, and so you’re not able to actually make real progress, you’re just working on demos and proof of concepts. And we didn’t even start with $1 million, we started with zero. So the first financing was I refinanced the house that I owned, which is comical when I think back now that my level of conviction and optimism. No, it’s awesome. RJS: I thought, “I’ll refinance my house, take the $100,000 to get out of it and use that to start a car company”, so that was what we did. But it’s very hard to then hire, we’re just getting a semblance of traction to have some capital, some amount of money that we could actually make real progress on a product. Right. And this was all — you still didn’t really know for sure what you were going to build, right? RJS: That’s what I was going to say, it’s actually really helpful that in those years we didn’t have capital, because we could have started building the wrong thing, so it provided me this few year period where I was learning how to run a company. I’d never run a company before, I was learning how to lead teams, I was learning how to hire, I was learning how to have hard conversations, I was learning how to raise capital, I was learning about strategy and design and brand and all these things. And so it was a really wonderful period of time for me because we were iterating so dramatically, so significantly on the strategy, the product, the type of company we’re building, the skillsets we want to accumulate and build in-house, in ways that we couldn’t today. I couldn’t walk in the door to Rivian, say, “All right, everybody, we’re going to do a completely different set of products, get ready”. You were on an e-bike back then, you could sort of go where you wanted to. The bigger you get, the more locked in you are. RJS: Yeah, the whole team could fit in one little room with one little table, investor management was very straightforward, it just gave us the freedom to be very iterative. And I look back and I’m so thankful that happened because this squiggly path led us to what Rivian ultimately became, this idea of building a really strong brand around enabling and inspiring adventure that scales across different price points and form factors. We came up with the concept for R1 as the flagship product, then we would follow that with R2 , which is now about to launch, and then R3 , which is going to launch shortly thereafter, and things took a lot longer. Once we even got all that defined, we still had to raise a lot more capital. We then raised a lot of capital and we’re on the path of execution, and there’s some big unplanned events. COVID was very, very, very challenging and maybe the worst possible time you could imagine it. Right. You’re just about to launch. RJS: Yeah, so trying to build a plant starting in 2020, which is sort of wild, and then turn on a supply chain with a bunch of suppliers that didn’t want to work with us, we had to pay extremely high prices to them just to get them to provide us components. Just a bunch of things that when you’re planning it years before, you don’t think, “Well, there’s going to be a supply chain crisis, there’s going to be a pandemic”, and there’s going to be all these externalities that make it really hard to start in that moment. You mentioned getting to this adventure brand identity, the R1T, R1S being your initial products, what was the process of honing in on that? Why did you decide that was the way to go? I mean, from the outside, you view obviously, Rivian, you’re always going to be compared to Tesla in a certain respect. They have this futuristic car looking very aerodynamic, and Rivian comes along, it’s like, “Yes, thank God, a pickup truck”, not a Cybertruck, they got an SUV. That’s my perception of the outside, but what was it like inside? RJS: Yeah, I mean, it wasn’t too different from that. We recognized that in order for us to earn the right to exist, we needed to do something that was unique and could stand on its own, and so some of the early things we thought about, we’d originally thought about doing a sports car, we realized that we were just going to be too close to what Tesla had done, and what Tesla had done well, by the way. So we went through deep soul searching to say, “What are the things we’re passionate about?”, “What are the things we want to enable?”, “What are the things that are going to matter?”, once everything’s electric, imagine every car on the road is electric, you can’t say we’re differentiated because we’re electric. “Why are you differentiated?”, “What is the reason for someone to choose to buy our products?”, and so we went through a lot of those thought processes and came out of it with this idea of preserving and inspiring the adventures that you want to have in your lifetime, the kinds of things you want to take photographs of. The reason why you want cars is you can go anywhere— RJS: Yeah, you can do all this, you can go to your grandparents’ house, you can go to the beach, you can go climbing. So that led to this really clear vision, which then led to product requirements. “Okay, if we want a car that’s going to enable and inspire adventure, what does it want to look like?”, “What are the features it needs to have?”, so storage becomes a really big consideration, being able to drive on any type of terrain becomes a big consideration, and then you say, “Okay, what’s the vehicle form factors that are going to do that?” — a re-imagination of a pickup truck and a re-imagination of a large SUV, that’s a great flagship. So it wasn’t always as direct as that single sentence, sometimes it took us a month to get there or more, but a lot of iteration, a lot of the product concepts, some of the early R1T stuff that we put together looked really futuristic and not inviting, which is the word we use all the time, like inviting you to use it. Not wanting to get dirty, or it didn’t want to get used or you don’t want to put a surfboard on the top so we became really very intentional around, “What is a Rivian?”, “What is not a Rivian?”, so we do all these exercises from a design aesthetic point of view, which of course now we know our aesthetic, but in 2015, we had no idea what a Rivian aesthetic was, so we had to define that. So we do these is/is not exercises, it’s all things that it’s amazing sitting here today to see it having played out where people actually connect and resonate with the brand that we were hoping to. It’s an incredibly strong brand, you can identify it right away. You mentioned the COVID production challenges, there’s also a bit where actually scaling up production is just really hard, even if everything is perfect. How do you distribute the blame between COVID and between the fact that actually, “This is much harder than I thought it would be”, in terms of the challenges in getting out the door? RJS: Yeah. I think we made a tactical or strategic error, which is we decided to launch three vehicles at the same time. Yep, that’s one of my questions coming up. RJS: Launching any vehicle is really hard. So just to put this into perspective, you have around 30,000 discrete components, which you purchase as a company as maybe 3,000 items and the reason there’s more discrete components is you buy something like a headlight as a single assembly, but it has many components in it. But all those tier two, tier three, tier four supply components, any one of those can stop production if they’re missing and so you still have to think about it considering the full complexity of all the parts, every single mechanical part that’s in the vehicle and so turning on a supply chain for the first time is hard for any product. For a car, it’s really hard. And for a car when the supply chain doesn’t want to work with you, meaning business is thriving, it’s very different than let’s say 2010 or ’11 or ’12 when the suppliers were all beat up from the recession and were willing to take any business. In 2020, they were busy, they didn’t want to take on this new customer Rivian with an unproven brand and unproven product. So it was very, very hard to get them to work with us and so just getting all those suppliers to ramp at the same rate on one car would’ve been tough. And the reason I say same rate, if some ramp faster than others and you have inventory issues. Right, then you have these working capital problems. RJS: You have to ramp at the same time so you can make a complete car and sell it, sounds so simple. (laughing) No, don’t worry. It does not sound simple at all. RJS: So we were doing that across three vehicles at the same time, that was already a big — the R1T, R1S launched at the same time plus a commercial van . And then on top of that, we had COVID, which made everything more challenging. Yeah, so it was maybe the most perfect of perfect storms for difficulty and so I wouldn’t use COVID as an excuse or I’m not putting blame there, I’m just saying it’s just a reality, it was just a reality. Oh, for sure. No question, I don’t think anyone is denying that. RJS: But if I were to do it over again, I would’ve launched probably the SUV first, spaced out maybe 12 months, then launched the truck, spaced out probably 12 months, launched the van, and had smoother launches that consume less capital that would allow us to get to profitability faster. But hey, you learn. And so here we are in R2- RJS: Yeah. So R2 is like, we’re launching one build combination, we’re launching a launch edition, we’re not launching R3 at the same time. You’ll laugh at this, Ben, there was a lot of debate like, “Oh man, R3 is so cool”, we had thousands of customers like, “Oh, we can’t wait to get an R3 as well, can you guys launch that quicker?”, and we’re like, “Should we try to launch R2 and R3?”, “No, no, don’t do it! Don’t do it! Hold it back!”, and so we held back. It’s like you needed to hire someone back in 2021 that says, “If we ever consider doing this again, stomp on the table and say ‘No’.” RJS: As a product person, you have all these ideas, you want to see them out in the world as fast as possible so simplicity and focus has been a major emphasis for us and so the entire business is laser locked on launching R2 and it’s a beautiful thing. We don’t have other programs that we have to manage, it’s like, “Let’s get that, that has to ramp quickly, that’s what’s going to drive us to profitability”. It’s key for cash flow, we have this enormous R&D spend we’ve created intentionally to build out all these vertically integrated technologies , whether it’s our chips, our software, our compute platforms, our high voltage architectures, that the scale that R&D necessitates the scale of more than just R1, more than just a flagship product that needs a mass market product, and that’s what R2 brings us. Got it. So you mentioned the van. Ideally from a production standpoint, you do SUV, then you do truck, then you do van. The van though came with a lot of money from Amazon , is that a critical component in why you launched it maybe sooner than you should have? RJS: No, at the time I didn’t think it was sooner. I mean, at the time I thought it was the right thing to launch them all at the same time. Amazon’s still our largest shareholder and they’ve been a great partner and they were an investor in us when we were private, as you said. But what’s so exciting about that program is it took a space that has the logistics based on last-mile e-commerce space that has such a clear value prop for electrification, meaning the vehicles start and end the day in the same spot, which is a great thing from a charging point of view. You know what they’re going to do, that you can deterministically control what they’re going to do in terms of mileage. You know your 99th percentile route in terms of number of miles, and you know your one percentile, so you can really optimize it for total cost of ownership, and so that’s what we did. So we went about and said, “Let’s make the ultimate delivery van, let’s make it the most cost-effective way to deliver”. So you talk about the complexity of doing three vehicles. Is that just in terms of getting started or is there a production capability, like you only do so many things, or is that part fine? It’s just the part of getting started? RJS: Part of the challenge is when you’re launching a manufacturing and supply chain infrastructure for the first time, in our case, we didn’t fully appreciate all the things you need to be really good at to do it and so we tried to very quickly learn how to be able to launch multiple programs at the same time, which eventually Rivian should be able to launch multiple vehicles in the same year at the same time, but we just didn’t have the maturity of process, maturity of our organization or the depth of teams to be able to support that. The issue is not things you can plan for, it’s all the little things you don’t plan for, and there’s all these little things, each of which requires problem solving. So I used to describe it in 2021 and 2022, is it’s not like there’s some giant unlock. Like, “If we just solve this, we will make more vehicles and we’ll get our cost structure in line”. There was just a stack of thousands, truly thousands of little things that needed to be adjusted or changed or negotiated and I think the thing that compounded all this that was really hard, is a lot of those issues were at our suppliers, and then those suppliers had a lot of leverage over us, because they know that in that time where they couldn’t get enough- If this doesn’t get done, you’re done. RJS: We broke up with a lot of these suppliers, but some of them would just say, “We want you to pay us twice what we previously negotiated if you want parts”, and we’d say, “No”. and they’d say, “Okay, fine, we just won’t send you the parts”, and we’d say, “Okay, how about one and a half?”. We just had no leverage. So that’s changed so dramatically and we see it with R2. R2’s the first, I’d say, clean sheet from a supply chain point. Even with the updates we made to R1, we were able to get rid of a lot of that and they call it inflation-related, COVID-related cost growth that was born out of a lack of leverage that we had, R2 is the first time we were able to really reset the negotiations. You think of it from the perspective of if you’re Volkswagen, the leverage is the other way around, which is Volkswagen has so much scale and so many diverse sets of suppliers that they could say, “Hey, if you don’t bring your costs down, we’re just going to switch to another supplier”, we didn’t have that. We couldn’t say, “Well, look, we’re going to pull this other program from you” — it was no leverage, so we sort of were complete takers in that. Yeah, that makes sense. Before we got into the AI stuff, I did want to ask about the VW partnership . This includes access to your electric vehicle software, electrical architectures, you get supply chain expertise from them. How do you characterize this deal as a whole? I should also mention a sort of massive investment on their side as well, give me the framework of that deal and why it’s important. RJS: Yeah, it’s a $5.8 billion deal, some of which is technology licensing, some of which are investments. Right, I was going to ask about that. Some of it is just actually putting money in the company, and some of it is they’re going to license your software and things like that going forward. RJS: Yeah, and a lot of those are upfront licensing fees, most of which have already been paid and before I get to the business of it, it’s important to talk about the mission of it. We’ve spent a lot of time developing what we call a zonal architecture, but essentially think of it as a number of computers consolidating into one that perform a wide array of functions across a physical zone of the vehicle and it allows us to do things like over-the-air updates very seamlessly because rather than having a bunch of smaller function or domain-based electronic control units, little mini computers run the software for different functions, we run all this software for those functions on one computer on our OS, which makes it much easier to update. And so the strategy there was, “Boy, we’ve spent a mountain of investment building this tech stack, it’d be really nice to see it applied in another way. Yup. You need to get leverage on that investment and you just don’t have the volume by yourself. RJS: And it aligns to our mission in terms of enabling more electric vehicles to get highly compelling electric vehicles on the road and then it gives us a lot of scale, scale for sourcing the components that are shared and then it gives us the benefits of other, what we think of as joint sourcing agreements, so sourcing partnerships that can exist with Volkswagen. It’s been a great relationship, those types of relationships are very, very hard to build because it does require buy-in from the top so one of the things that allowed us to work so well with Amazon, I mean, you think about Amazon and it’s one of the largest companies in the world, certainly the largest e-commerce company in the world, and imagine they go out and say, “We’re going to build our future logistics network around a van that’s being not dual sourced, but single sourced to one company” — this is in 2019 — “has never built a car before at scale, and they’re like a startup”. But that was born out of a great relationship that I had with [Former Amazon CEO] Jeff [Bezos] and Jeff’s trust in supporting us and that enabled them to really lean in with us and lean in in defining the product, defining what it was, that was a really big leap. So we’ve built, I’d say, organizationally, really great capability of taking the strengths of being a fast-moving startup and working with very large companies as partners and in the case of Volkswagen, my relationship with Oliver Blume , the CEO of the group — so Volkswagen Group is, we think of VW as a brand, but they’re a group — they’ve got Porsche, Audi, Lamborghini, SEAT, Škoda, these are brands that aren’t sold in the United States, but it’s the second-largest car company in the world, largest industrial company in Europe, a huge company. But having Oliver and I aligned just allowed us to really move through the deal mechanics and the deal structuring quite quickly. So this bit, as you sort of zoom out, the deal makes a lot of sense to me. Actually, I think it makes a lot of sense for both sides. RJS: Yeah, it’s a win-win. They get expertise that they’re not going to develop internally. I’ve had plenty of German cars, the software is okay for what it is, I don’t think it’s going to sort of go where you’re going to go. You also have on your side, you can do these huge investments like you talked about last week , and we’re about to transition into that, with the promise of scale that is much more than you can certainly deliver today. Is there a bit of you though is like, “If we had ramped up correctly, if we had not done multiple vehicles, we could actually be at scale, we could keep this all to ourselves”, or is this ultimately the best outcome that you’re sharing with them in the long run? RJS: I think in hindsight, I wish we’d ramped up more quickly, there’s things you’d change, but they’re also all things you’d learn from. We don’t spend any time lamenting them or anything like that. But to be clear, both in the case of our in house software and zonal controllers, which is what we’ve done with in Infotainment, which is what we’ve done with Volkswagen, as well as our autonomy platform and AI platforms , which is separate from the Volkswagen venture. Is that part of the deal? RJS: No, that’s not part of the deal. RJS: That’s 100% Rivian. RJS: But in both cases, we developed them thinking that we would eventually leverage this, not just with our own products, but with other companies as well. Got it. Okay. RJS: And so Volkswagen was, in many ways, the ideal first customer. And the reason I say it’s the ideal first customer, 1) it’s huge, as we’ve already described, but 2) it has the complexity of managing across many different brands, and so being able to support a company like Volkswagen Group, which spans very premium brands, like Porsche, down to one of the products that’s been announced that we’re doing together, the Volkswagen ID1 , which is a $22,000 EV, it’s the existence proof that we, Rivian, can support working across large complex organizations, across large ranges of price and product features, and across very different vehicle form factors. And if you’re another car company, you couldn’t look at Rivian and say, maybe before you could have, but now you couldn’t, say, “Well, I don’t think you could do this at this price point” — well, actually we cover every price point across the spectrum. So there’s an opportunity for other car companies to do the same thing. RJS: Absolutely, yeah. And now on the autonomy front, I think the opportunity there is actually bigger because this is a very, very hard problem to solve, it requires vertical integration in ways that are not typically — it’s just things that OEMs typically don’t do. Tell me your vertical integration story, because it is really interesting. You’re on last week, you talk about everything from your chip to your sensors to your software, you talked about building your own compiler. We are talking total front-to-back, end-to-end vertical integration. Why is that important? RJS: Yeah. It’s important to just talk about how autonomy is now being developed, and I do think for anyone listening to this, it’s very, very important to understand this because there’s perhaps some histories to how it was done before. The idea of a vehicle driving itself isn’t a new idea, that’s been something, it’s been in sci-fi movies for decades, but in terms of actual technology development, it started in, call it early 2010s, in that time range, so roughly 15, 20 years ago. The early platforms and what was done in terms of the approach up until the very early 2020s was something that was designed around a rules-based approach and so what you would have is you’d have a set of sensors, perception that would identify objects in the world, so all the things in the scene, so that’s cars, people, bikes, kids, balls bouncing on the street, everything that you can see, it would identify all those objects, it would classify the objects as to what they are, it would then associate vectors to those objects, acceleration of velocity, and it would hand all those objects and their classifications and their vector associations to a rules-based planner. The rules-based planner was a team of software developers attempt to codify what are the rules of the road. So, I’m going to oversimplify here, but think of it as a whole series of if/then statements. Totally deterministic, by the biggest spaghetti code mess you’ve ever seen because there’s so many possibile exceptions and issues. RJS: It’s a giant, giant code base that’s trying to describe how the world works. And so, it wasn’t actually AI as we think about AI today, there was machine vision. Machine learning, neural nets, yeah. RJS: Yeah, there was machine vision for the object detection classification, but in terms of the planning and the actuation of vehicle was very much a rules-based environment. Then along came the idea of neural nets, and the idea of transformers to do encoding, and that happened, of course, in the LLM world, but that’s also happening in the physical world. Everything can be a token. We think about it, everyone thinks one of the context of letters and words, but everything can be a token. RJS: Yeah, everything can be tokenized and the whole world changed in self-driving, so everything that was done prior to, call it 2020, 2021 is largely throwaway, meaning the way the systems are now developed is you build, you need to have complete vertical control, it needs to be one developer that controls all the perception, because you don’t want a pre-processed set of outputs from a camera, you want the raw signals from a camera. If you have other modalities like a radar or LiDAR, you want the raw signals from those, you want to feed it in through a transformer-based encoding process early, so fuse all that information early, and build a complex, it’s hard to imagine in our human brains, but it’s a complex multidimensional neural net that describes how the vehicle drives. Then you want to train that with lots and lots of data, and you’re training it offline. The word that gets used all the time is end-to-end, so it’s trained end-to-end from the vehicle through the human drivers back to the model and so, to do that well, you need a few ingredients, you need this vertically-controlled perception platform, you need a really robust onboard data infrastructure that can both trigger interesting data events, hold them, do something to them to make them a little easier to move off the vehicle, ideally through Wi-Fi, and a worst case through LTE, but mostly through Wi-Fi, all that data gets moved off the vehicles, and this is happening at millions and millions of miles accumulating just in the course of a day. And so all that data is moving off the vehicle, and then you’re training it on thousands and thousands of GPUs. You’re going around and around and around, and it gets better and better and better. That’s an approach that is so different, as I said from what was done before, but to do that, you need all those ingredients. Well, you need cars on the road. RJS: You need cars on the road. So, we looked at it, we launched in 2021 with our Gen 1 architecture, we almost immediately after that realized we needed a complete rethink of our self-driving approach. Right, that’s exactly what I was going to ask. Was this an issue where in some respects you launched later than you wanted to because all the supply chain issues, but then you actually launched earlier than you wanted to because you didn’t have the right sort of stuff on your cars? RJS: Well, we launched — and we didn’t realize, and this is the thing, and even some of our Gen 1 customers are not happy with this, but when we developed the Gen 1 system, this was on 2018, 2019, we didn’t know this big technical massive shift was going to happen. So, our Gen 1 architecture uses a Mobileye front-facing camera, and it uses — it’s a collection of things, it’s very classical rules-based approach, if you’re going to develop something around AI, it’s a completely different architecture, not a single shared line of code, not a single shared piece of hardware. So we started working in the beginning of 2021, right after launching on a whole new clean sheet, everything new, we didn’t try to morph anything over, it’s a complete melt and re-pour. In that new architecture, we designed cameras, we designed a new radar, we designed a new compute platform, we built, we call this our Gen 2 architecture. We built it around an Nvidia processor, we designed a data flywheel, we designed an offline training program. The vehicle launched in the middle of 2024, the features then were trained on a very small number of miles, which was our own internal fleet and now over the course of last year, we’ve built up enough data that’s allowed us this flywheel starting to spin. Yep. And that data is only coming from the Gen 2 vehicles, right? Not from the Gen 1 ones? RJS: Only Gen 2. Gen 1, it’s asymptoted, both in terms of capability and it has no value to us in terms of data, so only Gen 2. And so, in parallel to kicking off this Gen 2 platform, which we said, we need to get this in the field as fast as possible because we need to start the data flywheel, we also need to get better hardware so that when we have the model built, we can run it with a higher ceiling. That kicked off updates to the cameras that are going to our Gen 3 architecture, very importantly, an in-house silicon program. Why is that very important? RJS: Compute inference on the vehicle, we wanted to have — what would we have in our Gen 2 is around 200 TOPS [Trillions of Operations per Second], we wanted that to be closer to 200 TOPS per chip, so 400 TOPS total, sparse TOPS. Well, what’s going to be in Gen 3 will be 1,600 sparse TOPS, but importantly, we designed it specifically around a vision-based robotic platform. And so, the utilization of those TOPS is very high, much higher than what we see in other platforms that are more generalized, and then the power efficiency is very high, and then the cost is much lower. So we have a very, very high capability, low cost platform for which we can afford to put enormous compute in. All that is true, but the actual development of that is very expensive. Is this going to pay off with those lower unit costs, and that increased capability with just your vehicles, or like the VW deal, is this something that you’re going to be looking to sell broadly? RJS: Well, this is an interesting one. Even on its own within Rivian, just R1 and R2, it’ll pay off because the cost savings are so significant on the chips. But more than that, we believe we’re very, very — we’re spending billions of dollars in developing our self-driving platform, our level of conviction as this being one of the most important, I shouldn’t even say one of the most, the most important shift in transportation and transportation technology means that we want it to control the whole platform. Then once we control the whole platform, it makes it a very interesting system that can be provided to other manufacturers. And so, I think in time, the number of companies that will have all the ingredients to do what I’ve just described, they’d be very limited, I think there’ll be less than five in the West. Did you get any of this thinking from Jeff Bezos? Because there is a bit here where our cars are where we get out and develop this and prove it out, but the real payoff is to do the platform at scale across other entities. It sounds a little Amazon-like. RJS: Amazon’s our largest shareholder, and Jeff’s somebody I look to for a lot of inspiration on these kinds of things. So, certainly I think there’s some of that. We think of our vehicles as our own dog food, but we’re going to make a platform that’s so darn good that we think others will- You’ll sell a lot of vehicles. RJS: And if others aren’t buying our platform, we’ll monetize it through selling more vehicles, and we’ll grab market share. I think on both sides of that, we can win. I do think that it’s going to move far faster than anyone realizes. I think, the way I describe it is if you look at the last three or four years of development in autonomy, and you try to draw a line to represent the slope of improvement, and you look at the next three or four years, the two lines are completely unrelated. Totally agree. RJS: But the acceleration is going to be so fast. And what I’m surprised is people aren’t — I say this, I don’t think people fully realize it, but the LLM space should teach us that. Yeah. GPT-1, GPT-2, GPT-3, GPT-4. RJS: But look at the 1.0 architectures. Oh, which are rules-based. Yeah, to your real point, it’s exactly what it is. RJS: Rules-based. And look at the progress that was made on Alexa, let’s say, relative to the progress that’s happened on GPT-3, 4 now, and beyond, it’s just like they’re not even closely related. And so the same thing is happening in the physical world with cars, and if you don’t have a data flywheel approach, you’re just not in the game and there’s no way you can compete. And so, very few people have that, far fewer I think is right. A big differentiator between what you’re doing and what Tesla is doing, and we have to sort of come back to it, they shifted to the pure neural network approach, but they’re doing vision only. Do you just think that’s a fundamentally flawed decision? RJS: We have a different point of view. Right. Because you have radar and LiDAR too, is the difference there. RJS: Yeah. There’s a lot of alignment, and we both agree, and we’re both approaching it as building a neural net. So, I want to call that out that we have a very aligned view. Right. Your core philosophy is absolutely the same. And I think there’s an extent where Waymo is getting there as well. RJS: The same philosophy. And then it’s like, “How can we teach the brain as fast as possible?” is our question. They have the biggest fleet of data acquisition in the world, they have fewer cameras, that have far less dynamic range. When I say dynamic range, I mean performance on very low light conditions, and very bright light conditions. Right, yep. RJS: We have much better dynamic range that of course adds bill of material cost, but we did that intentionally. And then, we have the benefit of our whole fleet, all Gen 3 R2s, think of those as ground truth vehicles. They’ll have LiDAR and radar on them. Tesla just has a few ground truth vehicles that do have radar and LiDAR, but they’re trying to service the whole fleet. RJS: Yeah, I’m looking out the window here at El Camino and you just have to stand at the corner and see Teslas driving around and around everywhere. One will go by eventually, yeah. So that’s the question, is the benefit of putting radar and LiDAR on all your cars, is that just something you need to do now so you can just gather that much more data that much more quickly? Or is that going to be a necessary component for at scale, everyone has an autonomous vehicle and they need to have radar and LiDAR? RJS: Yeah, I think, the way I look at it is, in the absolute fullness of time, I think the sensor set will continue to evolve. But in the process of building the models and until cameras can become meaningfully better, there’s very low cost, very fast ways to supplement the cameras that solve their weaknesses. So seeing through fog we can solve with a radar, seeing through dense snow or rain we can solve with a radar, seeing extremely far distances well beyond that of a camera or human eye, we can solve that with a LiDAR, our LiDAR is 900 feet. And then the benefit of having that data set from the radar and the LiDAR is you can more quickly train the cameras. The cameras, when I say train, it doesn’t mean we’re in there writing code to do this. I think my audience broadly gets how this works, yeah. RJS: The model understands this and so you feed this in and the neural net understands because you have the benefits of these non-overlapping modalities that have different strengths and weaknesses to identify, “Is that blurry thing out there actually a car?”, “Is it a person?”, “Is it a reflection off of a building?”, and when you have the benefit of radar and the benefit of LiDAR, that blurry thing way off in the distance that the camera sees starts to become — you can ground truth that much faster. And then you teach your camera to figure out what it is. RJS: Then your cameras become better, and so that’s our thesis. And of course, that’s important that we have a thesis that’s different than Tesla, if we had an identical thesis to Tesla on perception- They just have way more cars out there. RJS: Yeah, the only way to catch up is with building a fleet of millions of vehicles, we want to catch up faster than that. So is it also sort of this advantage that — to what extent do you feel the auto industry, you start out and you’re sort of the outsider, you can’t get suppliers to help you, they’re ripping you off, all the sorts of problems you talked about. Now you’re like, I can imagine Volkswagen at a minimum is looking at you, “Please figure this out, we have a relationship, we can sort of jump on if need be” — do you get that sense more broadly from the industry? Because I don’t think anyone expects Tesla to share their technology, Google is sort of its own thing, do you have the potential to be the industry champion in some ways? RJS: We hope. I mean, I think every manufacturer has three choices, it’s pretty simple. They’re either going to develop their own autonomy platforms, they’re going to buy an autonomy platform, or they’re going to make this not a priority and they’re going to lose market share. But the last one, you have to accept that in not too much time, if you don’t prioritize this, you will lose market share. It’d be like trying to sell a house without electricity, it’s going to become so fundamental to the functioning of the vehicle. Why do you think that autonomy is so tightly tied to electrical vehicles? Because there’s no reason an ICE vehicle couldn’t be autonomous. RJS: No, no. It’s more coincidence, it’s funny. I’d say autonomy, connectivity and modern infotainment and electrification are all completely separate topics, so there’s no reason they have to converge into one thing. It’s more just coincidence that all these things happen to be occurring at the same time and the electric vehicles tend to be the more advanced vehicles because they’re on new architecture. So it’s why you start to see from other non-peer review manufacturers that their EVs tend to be the most advanced but autonomy doesn’t care if it’s an engine or if it’s an electric motor. Right. It makes sense that’s just how it happened historically. Right. It makes sense that’s just how it happened historically. I do need to ask this question, I think I know what the answer is, but people will be mad at me if I don’t ask. Why is there no CarPlay in Rivians? RJS: It is a good question, we get asked that a lot. We’re very convicted on this point. We believe that the aggregation of applications and the experience, and importantly now with AI acting as a web that’s integrating all these different applications into a singular experience where you can talk to the car and ask for things and where it has knowledge of the state of health of the vehicle, the state of charge, distance, outside temperature, everything becomes much more seamless in time if the vehicle is its own singular ecosystem versus having a window within the vehicle that’s into another ecosystem. And is that the issue, just the implementation effort on your side or that the customers are actually short-circuiting themselves? RJS: We could turn on CarPlay really quickly, but then you end up with — you either enter into the CarPlay environment and it’s like Apple’s, they get to play the role of aggregating what apps are there and how they decide what’s integrated, how it’s done, versus us, and I think where it becomes really important is when AI happens. Our view is a lot of the applications will start to go away and you’ll have your AI assistant. There may be things happening below agent to agent under the covers, but when you say, “Rivian, tell me what’s on my schedule for later today”, you don’t care that it has to go agent to agent to Google Calendar to pull that out, you just want the information, that interface becomes really important, it becomes so fundamental to the user experience and the whole user journey. So as we’ve thought about this, inserting any sort of abstraction layer or aggregation layer that’s not our own just is extremely risky and you start to build dependencies on that that are hard to reverse. Is there a bit where Tesla covered for you because they don’t have CarPlay either, but now there’s a rumor they might add it and it might make it a little more difficult to hold your convictions? RJS: Maybe. As it’s always the case on these things, I think there’s people that are really used to having CarPlay and our goal is to make it such that the car is so good that they don’t even think of that. And if they were to go back to CarPlay, they’d miss having the integrated holistic experience that we can create. It’s interesting because I thought you were just going to go more on the — you just gave this strong pitch for integration and top-to-down, side-to-side, that wasn’t the core to your answer. I think your answer made a lot of sense in the future best interface, I can see your customers getting themselves on a local maxima because that’s what they’re used to and it’s there and they’re missing how much better it can be. But I guess it goes to your point, infotainment and electrification and autonomy, those are all separate areas. RJS: So think of it like this. The challenge is CarPlay is not everything, so if you have CarPlay and the vehicle’s driving itself, in most CarPlay instances, it takes over the whole screen. RJS: There are instances where you could have a screen in a screen, but then that is very — I always joke, this is something Apple would never do. They would never have a screen in a screen on their own devices. They would say, we want to have one experience and so you have one screen that’s putting up information that’s very specific to the vehicle operation that are things that are like, “Is the door open or closed?”, and then you have another that’s mapping— It’s competing. RJS: It’s like you have two different UIs playing out and I just think it’s poor UI, it’s a poor user experience. The only reason people want that is they’ve been trained because they’re in cars that have such bad UI that the life raft to escape the horrible UI that is embedded in the car is CarPlay, and CarPlay is a really important function for that. If I’m in a non-Rivian or non-Tesla and I get in, it’s like a disaster and I’m like, “Oh thank goodness there’s CarPlay”. It has some thoughtful UI, but we have a really thoughtful UI and the few things that are missing we’ve been adding. So we brought Google Maps in, which was a big one, there’s more mapping platforms that’ll come in over time. We’ve got all the music platforms, including Apple Music, natively integrated. But soon with AI integration, I just think a lot of this fades away because you want a singular layer and that may mean we’re running ChatGPT to do some portions, we may be running Gemini to do other portions, but we get to be the arbiter of all this stuff under the surface. What are we using for onboard diagnostics? What are we using for on the edge knowledge? What are we using for cloud knowledge? All that we get to build and decide on ourselves. And I think importantly, given how fast the models are moving, we have the ability to plug or unplug different models at our discretion, we can decide what’s the best model to use. For the record, I agree with your decision. And I think if Tesla added CarPlay it would be a bad decision. And the reason is, I think unless you own one of these vehicles, I have a Tesla, I don’t have a Rivian, but the tangible difference is, and people say this, but until you experience it it’s not quite clear, it is a computer on wheels, and the way I think about it is for ICE cars that I’ve had, automatic windshield wiping is like a luxury feature or automatic lights. If you step back it’s like, “Wait, this is a software thing that we can do it once and do it generally, of course even your cheapest Tesla is going to have this feature and then you get to remove the physical control and you should never even need to interface with that”. And if your car is a computer first and foremost, you have to go in on the user interface, it’s nuts to put something else there, even if people are crapping about it in the short run. So there’s my pitch for you for that answer next time. RJS: And I also think that people that are in Teslas and Rivians that are actually driving it, the number of people that actually complain about it is very, very low. The number of people that say they’re not buying Rivian because of CarPlay is a higher number, but once you get into it, you’re like, “Oh, what was I worried about? This is really good!”, and I think the same trend exists for Tesla. Yeah. RJ, it was very good to talk to you, thanks for coming on, I’m excited to see how this develops. RJS: Yeah, this has been great. Thanks so much. I appreciate the time, Ben. This Daily Update Interview is also available as a podcast. To receive it in your podcast player, visit Stratechery . The Daily Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly. Thanks for being a supporter, and have a great day!
Disney made a deal with OpenAI, which both speaks to the durability of Disney's assets and to OpenAI's competition with Google.
Welcome back to This Week in Stratechery! As a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle; highlighted links are free for everyone . Additionally, you have complete control over what we send to you. If you don’t want to receive This Week in Stratechery emails (there is no podcast), please uncheck the box in your delivery settings . On that note, here were a few of our favorites this week. This week’s Stratechery video is on Google, Nvidia, and OpenAI . Why Does Netflix Want Warner? There are an entire category of stories that are shocking but, after a few moments, not surprising; Netflix buying one of Hollywood’s most iconic studios was not necessarily shocking — it’s been rumored for a few months — but it is surprising. Netflix dominates paid streaming distribution; why do they need to get into production as well? Both Andrew and I offer our theories of the case on Stratechery and Sharp Text , and we debated the same on this week’s episode of Sharp Tech . For me, the biggest answer is what’s becoming a theme: the specter of Google, in this case YouTube. — Ben Thompson And a Bit More Netflix. I love every Michael Nathanson interview on Stratechery, because the conversations are equal parts substance and chummy chemistry as two old friends size up the media landscape. Needless to say, this was a very good week for an extended conversation about the entertainment business , and I was delighted to see the transcript land in my inbox on Sunday night (yes, I get advance copies). Come for ribbing about previous long-running Netflix debates, and stay for two friends grappling with the logic of the deal for Netflix, regulatory questions to come, and the implications for show business. The interview is as timely this weekend as it was earlier this week, as all the questions surrounding this deal remain very much unresolved! — Andrew Sharp All About Flighty. If you’re a Stratechery reader or Sharp Tech listener, you’re probably familiar with Flighty , a flight-tracking app that Ben finds an excuse to recommend at least once a month. This week Ben interviewed the Flighty CEO, Ryan Jones , and we got the full backstory on how Jones went from the oil industry to Apple, how the Flighty app came to exist, and what its future looks like in the modern app environment. The interview is a fun conversation between two nerds who like building things, but more than that, the Flighty story is worth appreciating as a reminder of what tech can be at its best: a business identifies a problem, uses technology to fix it, and makes life better for everyone. — AS Netflix and the Hollywood End Game — Netflix is driving the Hollywood end game, likely confident it can increase the value of IP, and fend off YouTube. An Emergency Interview with Michael Nathanson About Netflix’s Acquisition of Warner Bros. — An interview with MoffettNathanson’s Michael Nathanson about Netflix’s acquisition of Warner Bros. and the Hollywood end game. Trump Allows H200 Sales to China, The Sliding Scale, A Good Decision — The Trump administration has effectively unwound the Biden era chip controls by selling the H200 to China; I agree with the decision, which is a return to longstanding U.S. policy. An Interview with Ryan Jones About Flighty and Building Apps in 2025 — An interview with Ryan Jones about Flighty, my favorite iOS app, and how the App Store has evolved over the last 15 years. Netflix and the Flattening of Everything — Whether the $72 billion Warner Brothers deal closes or not, the era of Netflix as big tech Switzlerland is now over. Netflix Buys Warner Bros. State Department Serifs Legends of the RISC Wars From Wheat to Cherries in Chile Trump’s Plan to Sell Advanced Chips to China; U.S. Concessions Piling Up Amid a Push for ‘Stability’; Macron and the EU Conundrum A Cup Week Mailbag: LeStreak, The Pat Spencer Revolution, Bucks PhDs, SVG on Jokic, and Lots More Netflix Opportunities and Anxieties, Merger Hurdles to Come, Hollywood’s Endgame and What Comes Next
An interview with Ryan Jones about Flighty, my favorite iOS app, and how the App Store has evolved over the last 15 years.
The Trump administration has effectively unwound the Biden era chip controls by selling the H200 to China; I agree with the decision, which is a return to longstanding U.S. policy.
An interview with MoffettNathanson's Michael Nathanson about Netflix's acquisition of Warner Bros. and the Hollywood end game.
Listen to this post : Warner Bros. started with distribution. Just after the turn of the century, Harry, Albert, Sam, and Jack Warner bought a second-hand projector and started showing short films in Ohio and Pennsylvania mining towns; in 1907 they bought their first permanent theater in New Castle, Pennsylvania. Around the same time the brothers also began distributing films to other theaters, and in 1908 began producing their own movies in California. In 1923 the brothers formally incorporated as Warner Bros. Pictures, Inc., becoming one of the five major Hollywood Studios. What the brothers realized early on was that distribution just wasn’t a very good business: you had to maintain the theater and find films to show, and your profit was capped by your capacity, which you had to work diligently to fill out; after all, every empty seat in a showing was potential revenue that disappeared forever. What was far more lucrative was making the films shown in those theaters: you could film a movie once and make money on it again and again. In this Hollywood was the tech industry before there was a tech industry, which is to say the studios were the industry that focused its investment on large-up-front costs that could be leveraged repeatedly to make money. Granted, Warner Bros., along with the rest of Hollywood, did come to own large theater chains as well as part of fully integrated companies, but when the Supreme Court, with 1948’s Paramount decrees, forced them to split, it was the theaters that got spun out: making content was simply a much better business than distributing it. That business only got better over time. First, television provided an expansive new licensing opportunity for films and eventually TV shows; not only were there more televisions than theaters, but they were accessible at all hours in the home. Then, home video added a new window: movies could not only make money in theaters and on TV, but there were entirely new opportunities to rent and sell recordings. The real bonanza, however, was the cable bundle: now, instead of needing to earn discrete revenue, the majority of Hollywood revenue became a de facto annuity, as 90% of households paid an ever increasing amount of money every month to have access to a universe of content they mostly didn’t watch. Netflix, which was founded in 1997, also started with distribution, specifically of DVDs-by-mail; the streaming service that the company is known for today launched in 2007, 100 years after the Warner brothers bought their theater. The differences were profound: because Netflix was on the Internet, it was available literally everywhere; there were no seats to clean or projectors to maintain, and every incremental customer was profit. More importantly, the number of potential customers was, at least in theory, the entire population of the world. That, in a nutshell, is why the Internet is different : you can, from day one, reach anyone, with zero marginal cost. Netflix did, over time, like Warner Bros. before them, backwards integrate into producing their own content. Unlike Warner Bros., however, that content production was and has always only ever been in service of Netflix’s distribution. What Netflix has understood — and what Hollywood, Warner Bros. included, was far too slow to realize — is that because of the Internet distribution is even more scalable than content. The specifics of this are not obvious; after all, content is scarce and exclusive, while everyone can access the Internet. However, it’s precisely because everyone can access the Internet that there is an abundance of content, far too much for anyone to consume; this gives power to Aggregators who sort that content on consumers’ behalf, delivering a satisfying user experience. Consumers flock to the Aggregator, which makes the Aggregator attractive to suppliers, giving them more content, which attracts more consumers, all in a virtuous cycle. Over time the largest Aggregators gain overwhelming advantages in customer acquisition costs and simply don’t churn users; that is the ultimate source of their economic power. This is the lesson Hollywood studios have painfully learned over the last decade. As Netflix grew — and importantly, had a far more desirable stock multiple despite making inferior content — Hollywood studios wanted in on the game, and the multiple, and they were confident they would win because they had the content. Content is king, right? Well, it was, in a world of distribution limited by physical constraints; on the Internet, customer acquisition and churn mitigation in a world of infinite alternatives matters more, and that’s the advantage Netflix had, and that advantage has only grown. On Friday, Netflix announced it was buying Warner Bros.; from the Wall Street Journal : Netflix has agreed to buy Warner Bros. for $72 billion after the entertainment company splits its studios and HBO Max streaming business from its cable networks, a deal that would reshape the entertainment and media industry. The cash-and-stock transaction was announced Friday after the two sides entered into exclusive negotiations for the media company known for Superman and the Harry Potter movies, as well as hit TV shows such as “Friends.” The offer is valued at $27.75 per Warner Discovery share and has an enterprise value of roughly $82.7 billion. Rival Paramount, which sought to buy the entire company, including Warner’s cable networks, bid $30 per share all-cash for Warner Discovery, according to people familiar with the matter. Paramount is weighing its next move, which could involve pivoting to other potential acquisitions, people familiar with its plans said. Paramount’s bid, it should be noted, was for the entire Warner Bros. Discovery business, including the TV and cable networks that will be split off next year; Netflix is only buying the Warner Bros. part. The Puck reported that the stub Netflix is leaving behind is being valued at $5/share, which would mean that Netflix outbid Paramount. And, it should be noted, that Paramount money wouldn’t be from the actual business, which is valued at a mere $14 billion; new owner David Ellison is the son of Oracle founder Larry Ellison, who is worth $275 billion. Netflix, meanwhile, is worth $425 billion and generated $9 billion in cash flow over the last year. Absent family money this wouldn’t be anywhere close to a fair fight. That’s exactly what you would expect given Netflix’s position — and the most optimistic scenario I painted back in 2016 : Much of this analysis about the impact of subscriber numbers, growth rates, and churn apply to any SaaS company, but for Netflix the stakes are higher: the company has the potential to be an Aggregator , with the dominance and profits that follow from such a position. To review: Netflix has acquired users through, among other things, a superior TV viewing experience. That customer base has given the company the ability to secure suppliers, which improve the attractiveness of the company’s offerings to users, which gives Netflix even more power over suppliers. The most bullish outcome in this scenario is Netflix as not simply another cable channel with a unique delivery method, but as the only TV you need with all of the market dominance over suppliers that entails. The most obvious way that this scenario might have developed is that Netflix ends up being the only buyer for Hollywood suppliers, thanks to their ability to pay more by virtue of having the most customers; that is the nature of the company’s relationship with Sony , which had the foresight (and lack of lost TV network revenue to compensate for) to avoid the streaming wars and simply sell its content to the highest bidder. There are three specific properties I think of, however, that might be examples of what convinced Netflix it was worth simply buying one of the biggest suppliers entirely: With regards to KPop Demon Hunters , I wrote in an Update : How much of the struggle for original animation comes from the fact that no one goes to see movies on a lark anymore? Simply making it to the silver screen used to be the biggest hurdle; now that the theater is a destination — something you have to explicitly choose to do, instead of do on a Friday night by default — you need to actually sell, and that favors IP the audience is already familiar with. In fact, this is the most ironic capstone to Netflix’s rise and the misguided chase by studios seeking to replicate their success: the latter thought that content mattered most, but in truth great content — and again, KPop Demon Hunters is legitimately good — needs distribution and “free” access in the most convenient way possible to prove its worth. To put it another way, KPop Demon Hunters is succeeding on its own merits, but those merits only ever had a chance to matter because they were accessible on the largest streaming service. In short, I think that Netflix executives have become convinced that simply licensing shows is leaving money on the table: if Netflix is uniquely able to make IP more valuable, then the obvious answer is to own the IP. If the process of acquiring said IP helps force the long overdue consolidation of Hollywood studios, and takes a rival streamer off the board (and denies content to another rival), all the better. There are certainly obvious risks, and the price is high, but the argument is plausible. That phrase — “takes a rival streamer off the board” — also raises regulatory questions, and no industry gets more scrutiny than the media in this regard. That is sure to be the case for Netflix; from Bloomberg : US President Donald Trump raised potential antitrust concerns around Netflix Inc.’s planned $72 billion acquisition of Warner Bros. Discovery Inc., noting that the market share of the combined entity may pose problems. Trump’s comments, made as he arrived at the Kennedy Center for an event on Sunday, may spur concerns regulators will oppose the coupling of the world’s dominant streaming service with a Hollywood icon. The company faces a lengthy Justice Department review of a deal that would reshape the entertainment industry. “Well, that’s got to go through a process, and we’ll see what happens,” Trump said when asked about the deal, confirming he met Netflix co-Chief Executive Officer Ted Sarandos recently. “But it is a big market share. It could be a problem.” It’s important to note that the President does not have final say in the matter: President Trump directed the DOJ to oppose AT&T’s acquisition of Time Warner, but the DOJ lost in federal court , much to AT&T’s detriment. Indeed, the irony of mergers and regulatory review is that is that the success of the latter is often inversely correlated to the wisdom of the former: the AT&T deal for Time Warner never made much sense, which is directly related to why it (correctly) was approved. It would have been economically destructive for AT&T to, say, limit Time Warner content to its networks, so suing over that theoretical possibility was ultimately unsuccessful. This deal is more interesting. The complaint, if there ends up being one, will, as is so often the case, come down to market definition. If the market is defined extremely narrowly as subscription streaming services, then Netflix will have a harder time; if the market is defined as TV viewing broadly, then Netflix has a good defense: that definition includes linear TV, YouTube, etc., where Netflix’s share is both much smaller and also (correctly) includes their biggest threat (YouTube). That YouTube is Netflix’s biggest threat speaks to a broader point: because of the Internet there is no scarcity in terms of access to customers; it’s not as if there are a limited number of Internet packets, as there once were a limited number of TV channels. Everything is available to everyone, which means the only scarce resource is people’s time and attention. If this were the market definition — which is the market all of these companies actually care about — then the list of competitors expands beyond TV and YouTube to include social media and user-generated content broadly: TikTok, to take an extreme example, really is a Netflix competitor for the only scarce resource that is left. Ultimately, however, I think that everything Netflix does has to be framed in the context of the aforementioned YouTube threat. YouTube has not only long surpassed Netflix in consumer time spent generally, but also TV time specifically, and has done so with content it has acquired for free. That is very difficult to compete with in the long run: YouTube will always have more new content than anyone else. The one big advantage professionally-produced content has, however, is that it tends to be more evergreen and have higher re-watchability. That’s where we come back to the library: implicit in Netflix making library content more valuable is that library content has longevity in a way that YouTube content does not. That, by extension, may speak to why Netflix has decided to initiate the Hollywood end game now: the real threat to Hollywood isn’t (just) that the Internet made distribution free, favoring the Aggregators; it’s that technology has made it possible for anyone to create content, and the threat isn’t theoretical: it’s winning in the market. Netflix may be feared by the town, but everyone in Hollywood should fear the fact that anyone can be a creator much more. In 2019, Netflix launched Formula 1: Drive to Survive , which has been a massive success. The biggest upside recipient of that series, however, has not been Netflix, but Formula 1 owner Liberty Media. In 2018 Liberty Media offered the U.S. TV rights to ESPN for free; seven years later Apple signed a deal to broadcast Formula 1 for $150 million a year. That upside was largely generated by Netflix, who captured none of it. In 2023, NBCUniversal licensed Suits to Netflix, and the show, long since stuck in the Peacock backwater, suddenly became the hottest thing in streaming. Netflix didn’t pay much, because the deal wasn’t exclusive, but it was suddenly apparent to everyone that Netflix had a unique ability to increase the value of library content. In 2025, KPop Demon Hunters became a global phenomenon, and it’s difficult to see that happening absent the Netflix algorithm. First, it is in part a vertical merger, wherein a distributor is acquiring a supplier, which is generally approved. However, it seems likely that Netflix will, over time, make Warner Bros. content, particularly its vast libraries, exclusive to Netflix, instead of selling it to other distributors. This will be economically destructive in the short term, but it very well may be outweighed by the aforementioned increase in value that Netflix can drive to established IP, giving Netflix more pricing power over time (which will increase regulatory scrutiny). Second, it is also in part a horizontal merger, because Netflix is acquiring a rival streaming service, and presumably taking it off the market. Horizontal mergers get much more scrutiny, because the explicit outcome is to reduce competition. The frustrating point for Netflix is that the company probably doesn’t weigh this point that heavily: it’s difficult to see HBO Max providing incremental customers to Netflix, as most HBO Max customers are also Netflix customers. Indeed, Netflix may argue that they will, at least in the short to medium term, be providing consumers benefit by giving them the same content for a price that is actually lower, since you’re only paying for one service (although again, the long-term goal would be to increase pricing power).
Welcome back to This Week in Stratechery! As a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle; highlighted links are free for everyone . Additionally, you have complete control over what we send to you. If you don’t want to receive This Week in Stratechery emails (there is no podcast), please uncheck the box in your delivery settings . On that note, here were a few of our favorites this week. This week’s Stratechery video is on Robotaxis and Suburbia . What the Times Missed in Its David Sacks Story. On Sharp Text this week, I wrote about the commotion that ensued in tech and media after the New York Times profiled Trump Crypto and AI Czar, David Sacks, including an OpenAI-style outpouring of Sacks support, why the piece failed on its own terms, and an entirely different story that went unexplored. While the Times focused on the private interests that may benefit under Sacks’ watch, there are better questions about the public’s interest in leaning on someone like Sacks , and why the government might need Silicon Valley expertise as it confronts a variety of tech questions that have enormous implications for the future of the Western world. — Andrew Sharp Atlassian’s History and the Near Future. My favorite part of every Stratechery Interview is Ben’s “how did you get here?” question to first-time interview guests, and this week’s interview with Atlassian CEO Mike Cannon-Brookes is a terrific entry in the series. Come for the story of how a Qantas Frequent Flyer program eventually led to a $40 billion software business in Sydney, and stay for Cannon-Brookes on how his company is adapting to the AI era, as well as his take on “correct, but chronologically challenged” snake oil salesmen. Finally, as a rabid F1 fan, I’d be remiss if I didn’t recommend the end, where Cannon-Brookes expounds on Atlassian’s role sponsoring and helping to transform the once moribund Williams team (a story that can also be marketed to enterprises the world over). — AS Code Red at OpenAI. I have, for three years now — i.e. ever since ChatGPT took the world by storm in November 2022 — been convinced that we were witnessing the birth of the next great consumer tech company. Today, however, there are very legitimate reasons to be concerned that OpenAI is going to eventually succumb to the Google behemoth, just as Yahoo, Microsoft, Blackberry, and countless others have; I still want to believe that OpenAI can be an Aggregator, but they don’t have the business model to match, and that may be fatal. I summarized all of these feelings in this week’s episode of Sharp Tech , which covered both this week’s Article about OpenAI and Nvidia angst , and Tuesday’s Update about the bear case for OpenAI . — Ben Thompson Google, Nvidia, and OpenAI — OpenAI and Nvidia are both under threat from Google; I like OpenAI’s chances best, but they need an advertising model to beat Google as an Aggregator. OpenAI Code Red, AWS and Google Cloud Networking — OpenAI is declaring code red and doubling down on ChatGPT, highlighting the company’s bear case. Then, AWS makes it easier to run AI workloads on other clouds. AWS re:Invent, Agents for AWS, Nova Forge — AWS re:Invent sought to present AI solutions in the spirit of AWS’ original impact on startups; the real targets may be the startups from that era, not the current one. An Interview with Atlassian CEO Mike Cannon-Brookes About Atlassian and AI — An interview with Atlassian founder and CEO Mike Cannon-Brookes about building Atlassian and why he is optimistic about AI. The Forest the New York Times Missed Among the David Sacks Trees — The New York Times failed to support its David Sacks headline, and ignored better questions about the how U.S. devises modern tech policy. Google Looms Alan Dye Leaves Apple Let’s Break Down the 45nm Process Node A Quiet Chinese Mobile Giant in Africa Trump, Takaichi and a Game of Telephone; Japan Jawboning Continues; An Internet Governance Study Session; China Making Trade ‘Impossible’ Wolves and Cavs Concerns, The NBA Cup in Year 3, Questions on the Magic, Suns, Thunder and Raptors The Game of the Week, A Giannis Inc. Emergency Board Meeting, Chris Paul Gets Cut at 2 a.m. in Atlanta OpenAI Declares a ‘Code Red,’ Alan Dye Leaves Apple for Meta, Questions on Tranium 3, Substack, and F1
Good morning, This week’s Stratechery Interview is with Atlassian founder and CEO Mike Cannon-Brookes . Cannon-Brookes and Scott Farquhar — whom I interviewed in 2017 — founded Atlassian in 2002; their first product was Jira, a project and issue-tracking tool, followed by Confluence, a team collaboration platform. Atlassian, thanks in part to their location in Australia, pioneered several critical innovations, including downloadable software and a self-serve business model; over the ensuing two decades Atlassian has moved to the cloud and greatly expanded their offering, and is now leaning into AI. In this interview we discuss that entire journey, including Cannon-Brookes’ desire to not have a job, how the absence of venture capital shaped the company, and how the company’s go-to-market approach has evolved. We then dive into AI, including why Cannon-Brookes believes that there will be more developers doing more, and why Atlassian’s position in the enterprise lets them create compelling offerings. Finally we discuss Atlassian’s sponsorship of Williams, the F1 race team, and why Cannon-Brookes thinks they can both help Williams win and also accrue big benefits for Atlassian. To repeat a disclosure I have long made in my Ethics Statement , I did, in the earliest years of Stratechery, take on consulting work for a limited number of companies, including Atlassian. And, for what it’s worth, I’m also a huge F1 fan! Go Max. As a reminder, all Stratechery content, including interviews, is available as a podcast; click the link at the top of this email to add Stratechery to your podcast player. On to the Interview: This interview is lightly edited for content and clarity. Mike Cannon-Brooks, welcome to Stratechery. MCB: Thank you for having me, Ben. So this is admittedly a new experience for me, I’ve already interviewed the founder of Atlassian , but it wasn’t you. I’m of course referring to Scott [Farquhar] . That was eight years ago, actually, before I even had podcasts. It was very brief, but hey, like I said, new experiences. MCB: That’s true. That’s true. And you wrote a consulting paper for us in 2014! I was going to disclose, yes, in the very brief period where I did consulting work, you flew me down to Sydney for a week, I had a chance to learn a lot about Atlassian. And on a personal note, that consulting contract helped me a lot, that was when I was just starting. It’s funny how small the numbers seem in retrospect, but maybe that’s why I’ve shied away from writing about you too much over the years, because it meant a lot to me. So I appreciate it and there’s my disclosure for the interview. MCB: Thank you. It’s a good piece of work. Don’t forget, ironically, we started as a consulting and services business and then decided that software was a better business model, so I think you did the same thing. You went the scalability route instead of the consulting work via Sydney. Absolutely. I’m not doing anything that doesn’t scale anymore, but I did love visiting Sydney, so it was great. MCB: Still, we pulled out the old consulting paper you wrote for us in 2014. Why are we going to win, why are we going to lose, everything else, it was classic Ben work. Was it good? MCB: It’s pretty good! It’s interesting, I’d probably be embarrassed if I read it today. Anyhow, the good news is that since it’s the first time I’m interviewing you, I do get to do my favorite segment, which is learning more about you. Where did you grow up, but also, where were you born? I know they were different places. Then, how’d you get interested in technology and what’s your version of the Atlassian origin story? MCB: Sure, I feel like I’ve heard this question 1,000 times! Where to start? My dad was in banking, he joined the glorious institution that is Citibank today, from England. Parents are both from Cambridge and bounced around the world a lot as part of that job. Took the, “Hey, we need someone to go to this country”, and he was like, “I’ll take that”. So I was born in America, in a period I lived in New York. To be honest, lived there for three months before I moved to Taiwan. Really? Whoa. I didn’t know that. MCB: Yeah, in 1980 when it was very different than what it is today. Yeah. Were you saving that to drop that off me? I had no idea. I thought you went straight from America to Australia. MCB: I only just thought about it about 30 seconds ago, actually. No, I went to Taiwan for a few years, lived in Hong Kong for a few years, went to Australia for a few years. So how I got into technology is actually related because my parents were moving around so much, the logic was being English, that they would send us to English boarding schools and that would be a stable thing while they were moving once we got old enough. So at the mighty age of seven, I was put on Qantas and sent to England and back four times a year to go to boarding school in England for about five, six years. Because of that boarding school, I have one of the lowest frequent flyer numbers in Australia, they introduced the frequent flyer program and that was at the end of year one or end of year two. I get given this catalog by my parents and how you’ve earned all these points, “What do you want to buy?”, and it’s like, “I don’t know, trips, winery things, booze”, I’m flicking through this catalog and I’m like, “There’s literally nothing in this catalog”, of gear that you used to be able to get that I wanted and at the back is this computer, so I was like, “I guess I’ll get that”. The only thing that was potentially age appropriate. MCB: That was the only thing in the catalog, I didn’t want a toaster, I didn’t want wine, so that became my first computer, the mighty Amstrad PC20 . Four colors, no hard drive. Eventually, I bought an external floppy drive, so you could put in two and did buy magazines and type in programs and write games and stuff from magazines and play with it, played a lot of video games basically back in that era. I was into computers peripherally all through high school, came back to Australia at 12, my parents had settled here by then and weren’t moving, and so I came back here, did all high school and university here. In high school, I was always going to be an architect, that was my dream the entire way through, but come to the end of grade 12, applied for a bunch of scholarships, because university, applied for the scholarships, ended up getting one and so I thought, “Oh, well, maybe I’ll take that”, and it was in a course called BIT. Basically, half computer science, half finance and economics, but it was 15 grand a year, tax-free, so I was like, “Well, I’ll do that for a while and go back to the architecture thing”. Of course, famously in that scholarship, I met my first business partner of my first startup, met my second business partner of the second startup, they went in radically different directions in terms of outcome, but it was just 30 kids right at the right time, did the dot-com era thing. Now, ironically, as a part of that scholarship, you had to spend six months in three industrial placements, so the origin story of Atlassian comes from then a little bit, because those industrial placements were so boring. Scott spent six months installing Windows at a large corporate and he was crazy freaking smart and it was like, “Hey, go from computer to computer and upgrade to Windows 98”, or whatever it was. It was like, “Guys, this is our life, this is going to be horrible”. I worked for Nortel Bay Networks, which was a good, at the time, massive competitor, Cisco then completely disappeared and so a good tech lesson in and of itself, I basically cataloged the room full of networking gear and routers, it was mind-numbingly boring. So towards the end of the university course, I famously sent an email to a few people saying, “Look, I don’t really want to get a real job, why don’t we start a company and we’ll try some stuff?”. And this was after the dot-com era? This was the early 2000s? MCB: This was after the dot-com era, yeah. So I lived through the dot-com era actually as a journalist and writer, analyst and technology. I worked for a company called Internet.com, which became Jupiter Media and Jupiter Research and that was great, that was an amazing era for me. We ran events, newsletters, what would’ve been podcasts, didn’t have them back then. And we ran events on Mobile Monday, I think one of them was called and it was all about WAP and— Well, the real secret is you’re not the only one. There are some founders that are very successful, that they’re like, “Look, I just want to pontificate about technology”. MCB: A little bit like you, I remember getting in a lot of trouble from some of the startups, because some company would launch and I wrote basically 500 words on, “This thing’s never going to work, this is a disaster of an idea”, and they would ring up and yell at my boss and he was awesome, he’d be like, “Dude, just keep writing what you think”, and it didn’t make you very popular as a journalist type. Anyway, emailed some people, tried to start a business, we didn’t actually know what we were going to do. Atlassian has, I always tell people, a terrible origin story. You should not copy us. You just didn’t want to be installing Windows or upgrading software. MCB: We literally did not want to get a real job. And Scott replied and said, “Yeah, sure, I’m in for trying that”. He was one of the smartest kids in our class and his nickname is Skip, because he was the president of our student association and always a leader type and Eagle Scout and everything else, so we’re like, “Yeah, okay, let’s do that, we’re good mates” — and that started Atlassian. We picked the name in about five minutes, which if you consulted any branding company, would not have been chosen. Ironically, originally, we were going to do customer service and consulting, that was what the gig was. Hence the name, because Atlas was a Greek titan whose job was to stand on top of the Atlas Mountains and hold up the sky, that’s what he was supposed to be doing. He was a bad guy, so his punishment was to hold the sky up and we thought that was an act of legendary service, and so we were going to provide legendary service by holding up the sky for customers and as I said, did the service thing for about six months, decided that this is a terrible business. People paying us $350 US to answer their questions and didn’t scale and was at crazy hours of the morning and night and everything else. So in the meantime, we wrote the first version of what became Jira . We actually wrote three pieces of software, one was a knowledge basey type tool, one was a mail archiving tool for groups, so you could see each other’s email as a shared archiving. And were you seeing this and you were building tools for yourself, for your consulting business? MCB: Literally, yes, exactly. So all three were tools that we needed for ourselves. People would email us and I couldn’t see Scott’s email and he couldn’t see mine at the time and it was like this is silly, and we built Jira to handle questions and issues and problems that we were having ourselves that became a teeny bit popular. There was this glimmer that someone else cared, so we poured all the effort into that. What was that? What was the glimmer? Because this is when Agile is taking over software development and at least the legend is Jira and Agile go hand in hand, is that a correct characterization? MCB: A little bit, but this is actually pre-Agile. So Jira comes out before Agile is even a thing. I think it was about two or three years before we had any version of marketing or feature sets that involved Agile. This was just a web-based, at the time, a bug tracker. So the interesting evolution part of the company obviously is it started as a bug tracker for software developers, it became an issue tracker for technology teams and now it’s like a business workflow for tens of millions of people every day across the world, most of whom have nothing to do with technology, so it’s gone on its own evolution. Would anything have been different if this was the plan from the beginning, or did it have to be this organic, “We’re figuring it out as we go along as we’re running away from Windows installations”, sort of story? MCB: I think, look, obviously, if we could choose to follow in our own footsteps, the Back to the Future skeptic in me would say it’s gone pretty well, so I’d follow every single footstep I took. (laughing) Yep, totally. MCB: And that would’ve become the plan. But look, we had two hunches really, which both turned out to be radically correct. Now, I would say we were following waves or whatever else, but one was that the Internet would change software distribution, which sounds ridiculous now and when I talk to graduates nowadays, I have to put them in the right time and place and say, “Look, when we started, software was distributed on a CD”, BEA WebLogic was the bee’s knees and you used to have to get it on a CD if you were lucky. If not, someone would come and install it for you and that’s how software was distributed. We made that CD into a ZIP file and put it on the Internet for people to download. You didn’t access it like a SaaS application, you literally download it from our website. Right. It’s funny that when you first say that, it’s like, “Oh, it’s completely transformative”, well, but you were an on-premises software story. But actually, no, there’s several steps to getting to SaaS, one of which is just downloading software. MCB: And we had people call us before they would download to check that we were real and stuff and I’m like, “Why don’t you just download the damn ZIP file?”, and I also date them, because, well, maybe I’ll get to the business model part, but the second innovation was that we thought open source would change software costs. So we had this big hunch, we were both writing a bunch of open source code at the time. Open source was a massive movement, especially in the Java space. Embarrassingly, I actually wrote a book called Open Source Java Programming that you can find with some mates. It’s still on Amazon and we sold a few thousand copies, I think, but I swore I’d never write a book again, it was a very painful experience. Thank you, you’re validating my life decisions . MCB: Yeah. Open source did bring the cost of building software down radically. We were writing a very small layer, 5% of the code at best on top of masses of amazing open source libraries and we contributed to those libraries, but we could deliver an amazing experience for a very low cost. We learned a lot, pricing and packaging. So what was the implication of that hunch though? Just that the market for developers, that would subsequently mean there was more software? MCB: A little bit that was the implication of the hunch. Largely for us, it was that the cost was going down. Pre-open source, you had to write everything so if Jira was back then, I don’t know, a million lines of code, if you added all the open source libraries together, it was 25, 30, 40 million lines of code. It was so big that it was so expensive, because you had to write all of that. To think of Windows, they wrote everything, the networking stack, there were no libraries, there was no open source involved in the original versions, it was all written by Microsoft. So the cost of that was very high, then you had to charge a lot of money. So we thought, look, if we could take all these amazing open source libraries, contribute back to them — we were a great open source citizen — and build a piece of proprietary software on top of them that solved customer’s problems, we could deliver that really cheaply. In fact, we sold the original versions of Jira, they were $800, unlimited users, unlimited use with no lifespan. So it was just 800 bucks, one-time fee forever and we learned a lot about pricing and packaging firstly, but secondly, it was very simple. Our goal in the early days, we had to sell one copy a week to stay alive, that was it. Some weeks, we’d sell two copies. $1,600 US would roll in and we’d be like, “Cool, we got a week off to survive”, and then one copy a week became two and two became five and five became ten, and now it’s hundreds of thousands. Well, isn’t the thing you just didn’t want to have a job? So I love this part of the story, because when I started Stratechery, I had a job from Microsoft that made, I think, $104,000 or something like that. I’m like, “I just want to make that, because I don’t want to work for a corporation, so if I could just get to there, it’ll be great”. MCB: We had exactly the same sets of goals. We had a few things we wanted to make somewhere that we wanted to go to work. I wanted to get up every day and think, “I want to go to work”, and weirdly, almost 24 years later, I love coming to work, so a tick achieved. We wanted to make it so we didn’t have to wear a suit, neither of us really like wearing suits at all — in fact, it’s a bit of an allergic reaction often and so tick, don’t turn up to work in a suit every day. And thirdly, most of our friends, so this is right where IBM bought PwC ironically, so out of the 30-odd kids in our class, maybe 10 went to IBM as consultants and 10 went to PwC and then they all end up going to the same shop and their grad salary there was $47,600. So our goal for year one was to end the year making at least a grad salary and convince ourselves we’re not crazy kind of thing and we smashed that goal, so that was good, but that was there. The Internet, the distribution part is important, knowing your favorite topics. Tell me about that and along with the business model, because again, this goes back so far, I don’t think people appreciate the extent to this entire idea of self-serve or bottoms up selling. This is really where it all started. MCB: Yes. And look, a few things. Firstly, if you come from Australia, we’re an exporting nation. “We’re built on the sheep’s back”, is a phrase, Australia’s built on the sheep’s back. What that really means is because we were this colony originally, then country on the far side of the world, anything we did to make money largely had to leave the country and go somewhere else. Originally, that was a struggle to find a product that could do that. “Built on a sheep’s back” is because wool was the first product that could do that, you could put it on a wooden boat, because it wasn’t very heavy and you could ship it a long distance, because it kept really well, so we could make sheep’s wool and make money as a country by shipping it back to Europe and it could survive the journey and so the country was built on the sheep’s back. We are a massive exporting nation. Trump brings in his tariffs, we’re the only country with a negative rate of return, we have a positive trade relationship with America and we’re like, “Wait a second, why did we get taxed?”, so obviously, it’s rocks, technology, we build and export everything as a country that we do. So our mentality was like, “Well, if we’re going to make money, it’s going to be overseas”, that was the first thing, is, “Okay, it’s going to be somewhere else, it’s not going to be Australians buying our software”, and so the Internet allowed us to do this. We put up a shopfront, early website and people could come to our website, download our software and then we just needed a way to get paid for it. The problem was in order to do that and the trust barriers of the Internet, we had to have a very low price and we had to have a fully installable offering. So we spent so much time on making it installable, documentation, “How would you get yourself up and running and try it?” — the software, as we put it, had to sell itself. Our software had to be bought, not sold. We didn’t have any salespeople, we couldn’t travel to your office in Sweden or London and help you out with it. For $800, we couldn’t have done that and secondly, it didn’t make any sense. So the evolution was, “Okay, this is the only possible path that we can go down is we have to figure out how to get people to do this”, now it turns out once you have figured out how to do that, it’s an incredibly powerful motor because you have lots of people coming, you have a very cheap piece of software for its relative performance, and you get people using it in all these big businesses all over the place. I would say 50% of the customers I go meet nowadays, probably meet a handful of customers, a couple a day on an average kind of thing, many of those have been a customer for 20 years, 22 years, 23 years. How many customers have been a customer 23 years? I’m like that’s crazy, we’re only 24 years old. That’s awesome. MCB: And so they downloaded very early, they didn’t download as all of , all of them are customers. Just one guy who’s like, “I need a way to track my issues”. MCB: Exactly. It was some guy in a backroom who needed to track it. I know the Cisco origin story, that was literally a guy, he’s still there, he’s been there 22, 23 years, he’s awesome. And they started with just, “I just needed a way to manage my issues for 10 people”, and now it’s hundreds of thousands of people, seats that we have there, it’s kind of grown over time. How did we know that business model was working? Again, it dates us a lot, this didn’t mean we didn’t answer questions, we were big on customer service and helping people, email was the way to do that. A bit of IRC back then, we had a channel you could log into and we’d help you. But the first customer, we used to walk into the office in the morning and we had a fax machine with literally rolls of paper. So if you wanted to pay for this distributed software, this says how old, there was no SSL keys, I heard you complaining about it the other day, totally agree with that era. You had to download a PDF off our website, which was pretty modern that it was a PDF, fill in your credit card details, and fax it to us, that is how you paid when we started. So we would walk in the morning and there’d be these rolls of paper on the ground, you be like, “Ah, sweet, someone bought something”, you know what I mean? It became a weird dopamine drug for us. The very first company was American Airlines… MCB: About six months in that we came in the morning and there was a fax on the ground with $800 and a credit card number written on it and we had never talked to American Airlines, they had never emailed us, they had never asked for customer service, they’d never gone on IRC, they had never talked to us in any way, shape or form. Man, this thing could work, we just made $800 out of the air. MCB: I mean, there was a lot of pre-work to get them there, but obviously that was kind of different. MCB: Then secondarily, as you wrote, I’m just trying to finish a very long answer here, we started Confluence in 2004, and those two became the jewel engines and both of those I think were probably major moments. I often say Confluence is a bigger moment, actually. The business model was kind of established, this is two years into the business. We made, I think, $800 grand in year one, $1.6 million in year two, maybe $5 million in year three, and $12 million in year four, if I remember the revenue numbers. So the thing was working really well. You’re the company that’s the Microsoft heir in some respects, which is the really just you took venture eventually, but didn’t really need to, just pure bottoms up. You and Scott, we’re able to keep a huge portion of the company because of that, it’s an amazing story that is, I think, under-told in some respects. MCB: Yeah, well, we actually did. I mean, we did and didn’t. So the venture story is one of my favorites because it describes how we think from first principles. Firstly, the first capital we put on the balance sheet, institutional capital to put on the balance sheet, I guess you could argue our initial, I don’t know, $10 grand each was some money, but was in the IPO . So in 2015, when we went public, that was the first capital that went into the business all time. We took two rounds of funding, one in 2010 and one in 2013, but both of which were to employees, the first was to the founders and the second was to large number of employees who bought in so both of those companies bought ordinary stock. Secondary shares basically, yeah. MCB: They bought ordinary stock, there were no preferences, there were no anything, that was kind of the way it is. And we love the Accel guys that invested, it’s kind of funny because their business model was wildly wrong, we now have their original spreadsheets and stuff. We’ve 15 years in, you know them really, really well, they wanted us to grow it. I think we had to grow at 30% for two years, 20% the year after and something like that to double or triple their money and at the time they put in $60 mil US , that was the largest investment I think Accel had ever made in anything software, digital kind of world and it was this massive bet. It was a one-page term sheet for ordinary stock, so credit to those two partners who took massive risk on us, had to fight, we know that GC, everybody else to do this unusual funding round and I think we did 50% growth the first year, and our CAGR since then is probably 40%. Yeah, it worked out pretty well. MCB: They did very well. I think their 2-3x was more like a 300x or something. You mentioned the Confluence moment. Why was that a big deal? Usually the story is you have one product and you need to focus and you’re two years old, you’re launching a completely new product. Is that the aspect you’re referring to? MCB: Yes, I think it comes down to being bootstrapped. Look, we spent nine years convinced we were going to die every day, there was just such a mentality that this thing was all going to fall over and we better work harder and keep going. The Confluence moment was important because I remember, I don’t know exactly, but sometime around then we understood venture capital. Firstly, on the venture capital side, because they do relate to each other, there was no VC available in 2001 and 2002 in Australia. We’re a nuclear winter, we’re two idiots with no credibility. Right. You could barely get funded in San Francisco, you’re not going to get funding in Sydney. MCB: No, because 2001, you weren’t even finding San Francisco funding because the whole dot-com boom had just happened, no one was getting funded anyway. We’re in Australia and we have no credibility, so we didn’t even bother. We literally, 2010 when we went to the Accel thing and we talked to five VCs, was the first time we’d ever pitched the business. It was just not a thing, people don’t understand, we used to say we were customer-funded when people would ask the also awkward question of, “Who’s your funding come from?”, we were like, “We’re customer-funded”, They go, “Oh, okay”. Lifestyle business! MCB: But we did understand venture capital, massive readers, I have an army full of technical books, books about technology and the industry and history and stuff from that magic era of airport bookstores. We read every episode of Red Herring and Industry Standard and Wired Magazine, I have just this huge library, so voracious readers. One thing you understood about venture capital is they put the portfolio theory on their side — and I’m a big fan of venture capital, I should say, I’m the chair of Australia’s biggest VC fund and that’s my other mate that I met in university, Niki Scevak . But we wanted portfolio theory on our side, we’d done finance and economics, we had one product, this was highly risky if you’re bootstrapped. So there was a little bit of the thinking that actually if we have two products, our chances of total failure are less, one of them can fail and we’ll be okay and so we started a second product. Yes, arguably it was hard, but our first one was going all right, it was like making, I don’t know, five million bucks a year and we had a handful of really awesome backpacker programmers. And the early people, it’s like a whole total band of misfits that somehow made this thing work and we’re having a lot of fun, we’re working really hard and so we made another internal tool that became Confluence and being adjacent, but very different, selling to different audiences, but having a lot — if you bought one, there was a good reason to have the other one, no matter which way you started, became a really good symbiotic loop of these two engines that powered us for a very long time. So it was more a case of reducing our risk actually than anything else. Wasn’t it risky to be splitting your resources or did that not even occur to you? MCB: I don’t think it occurred to us, no. It was more about splitting our risk and we were doing pretty well, but it changed the business because we moved from being the Jira company to a software company, and I say that’s probably the most under-understood moment because we had to learn about not how to market Jira, but how to market software, not how to build Jira, but how to build software. So now we have 20, 25 apps in 5 different categories that sell to all sorts of different teams who own a business, but we had to become a software company. Microsoft, I don’t know the analogy’s really that fair to them, to be honest, or fair to us, it seems massively over-glamorizing what they’ve achieved, which is amazing, I’m huge fan of Microsoft. The need to understand how to sell, in their case, like Minecraft, SQL Server, Azure, AI, you have to understand the building, the creation of technology, the selling of technology, the marketing of technology at a generic level, it really helped us generify the business. I think if we’d gone too much longer, everybody would’ve been on the Jira team, it would’ve been too hard to start a second thing and instead, we’ve always been a multi-product company. You just mentioned selling a lot. When did you finally realize or transition away from just being self-serve to actually, “We’ve got to grow beyond this”? Was it almost like a pivot that came too late because your identity was so wrapped up into the, “We’re the self-serve company”? MCB: Look, it’s never been a pivot, I get asked this by investors all the time. I would say our go to-market model and our process has kept evolving pretty much every year or two for 20 years and I say evolving because we’re very aware of the strengths of the model that we came up with and we’re very aware of what it takes to power that and we’ve been very careful when we’ve evolved, changed, added to it, not to destroy the original one. So nowadays, we have two amazing business models where we call them high-touch and low-touch. So we have the low-touch model, which is literally the same thing as it’s always been, hundreds of thousands of people show up every week, they try our software, we want them to have a great experience trying the software, we want to spread it as widely as possible and as many enterprises as we can, and some of those will stick, some of those will get working and we measure aggressively the rates of return and dollars and flows and funnels and everything else. This whole team whose job is to make sure that that’s working at now massive scale, right. But at the same time, what happened is as customers got more and more Atlassian software deployed, they wanted a different relationship with us, they wanted a bigger relationship. Those days they used to be spending, as soon as we were spending $20 grand, we were like, “Oh man, maybe we should talk to these people”, nowadays it’s more like around $50 to $100 grand is when we’ll talk to you. So the lines kept moving for different reasons and we actually have online sales, inside sales in between actually, the sort of classical someone gets on an airplane and goes to travel to you. So it’s just kept evolving. We talk about the IPO a lot, it’s our 10-year anniversary coming up this month, I’m off to New York next week to ring the bell and celebrate 10 years. When we went public, as an example, we had less than 10 companies paying a million dollars a year, now we’re well north of 500 in 10 years. So that doesn’t come without an amazing enterprise sales team and teams that go out and help customers and customer success and all the trappings of a really top flight enterprise sales organization, because for most of those customers, again, I think it’s north of 85% of the Fortune 500 are deep Atlassian customers. We become a strategic partner to these businesses that if we go down, rockets don’t take off, banks shut down, it’s a real critical importance to most of these customers. How big is your business outside of directly working with developer teams? As I recall, this was part of the consulting thing was you were wanting to do Jira for sales or Jira for all these different sort of functions, where and how did that evolve? MCB: So it’s been a continuum for a long time. So nowadays, less than half of our users are in technology teams, and probably a third of those are developers, less than half of them. So a portion of our audience, it’s a very important point of words. When I talk about this, all the engineers are like, “Hey, you don’t care about us anymore”, I’m like, “No, that’s not true”, that business is a great business, it’s just the rest of our business has grown massively around it. There are not enough developers in the world for our business. Our fundamental value has always been actually, and it took us one of these things, it took a decade to realize, firstly, we don’t solve technology problems, we never have, we’ve never had anything that’s like, “I care what code you write, which language the code is in, what the code does”. We solve collaboration and people problems, we always have solved people problems, even Agile was a people problem. It’s not a technology problem, actually, it’s a people problem. It’s, “How do we organize a group of people to build a piece of technology that best meets the customer’s needs and goes off track as little as possible?”, that is a collaborative people problem, we’ve always solved people problems. Our value actually came because there’s a lot of tools for technology teams and we never wanted to be in the dev tools business, that’s a road of bones, it’s very hard to build sustainable competitive advantage and dev tools, the history shows this. There’s just a different company every few years, developers tastes are fickle, our developers taste are fickle, this is not me sledging developers at all, we have a massive R&D arm and that group changes languages every couple of years, they change how they build software every couple of years, they’re constantly moving on, they change our analytics tools and everything else because they are tool builders and toolmakers, that makes sense, but that’s a hard place to build a business. Interestingly topical today, so we’ll see. But the easier place to build a business in the long term was the level above that, which is the collaboration problems that came, which started as, “How do we get engineers, designers, product managers, business analysts to all be on the same page about what it is that they’re building and have a repeatable process for that?”. It turned out that as the world has become technology-driven, as we say, our customers are technology-driven organizations. If you’re a large organization for whom technology is your key distinct advantage, it doesn’t matter whether you’re making chips and databases or whether you’re making rockets or cars or whether you’re making financial services or insurance or healthcare, I would argue for most of the businesses that are great, technology is their key competitive advantage, then you should be our customer, that is it. And what we help you do is we help your technology teams and your business teams collaborate across that boundary because that’s actually the hardest boundary. Building great technology is one set of problems, making it work for your customers usually means in different industries, a different amount of working with all sorts of business people and that’s what Jira did from the very start. Now that’s what our whole portfolio in service management, in strategy and leadership teams is about doing that at different scales and different amounts in different places. Does it bug you when you get complaints on the Internet of, “Jira’s so complicated”, “Hard to use”, blah, blah, blah? And are you speaking to, the problem is that the problem space we’re working in is not the single developer trying to track an issue, it’s trying to herd a bunch of cats and get them the same direction and muddling through that is a lot more difficult than it seems. MCB: It bothers me anytime people don’t like our software, sure. We’ve worked for the last 20 years to make it better every day. We’ll probably work for the next 20 years to make it better every day and people will still probably be dissatisfied and that is our fundamental core design challenge. There’s a few reasons they say that. Firstly, the on-premise business model and the cloud shift is really important because with the cloud shift, we update the software, with the on-premise business model, we don’t, so you would often be on older data versions, customers would upgrade once a year or every two years or something, and so we can’t control that. Secondly, the challenge of Jira is at our core, we solve a whole lot of what we say is structured and unstructured workflows. Confluence is an unstructured workflow, Jira’s a very structured workflow. You have a set of steps, you have permissioning and restrictions, you have fields, you have what’s happening in this process. The auditor will do something and pass it to the internal accounting team, the accounting team will do this and pass it to legal, legal will do this and pass it to these people. You’re defining a workflow and you’re having information flow back and forth and a Jira work item is, as we call it, it’s a human reference to work. That’s the best description of what Jira is work in the knowledge work era is this very ephemeral concept. Back to your development example, is the code the software? Is the idea the software? Is the designs in Figma — these are all parts of what it is, this thing that’s called this virtual thing that we’ve built. What we track is with a human reference to that, so someone can say it’s a new admin console. Cool, here’s the design for the admin console, there’s the spec for the admin console, there’s the code for the admin console, here’s where it’s been tested, here’s where it’s deployed. Did customers like it? We need a reference to this thing that is otherwise spread across hundreds of systems and virtualized. Once you’re building a workflow system, companies, ours included, love process, we love workflows, we love control, and that control usually comes with more data. “Hey, don’t fill in these three fields, fill in these 50 fields”, and they’re all required for some reason and our job to customers is to say, “Do you really need 50 fields?”, because you’re creating a user experience- You’re ruining it for us! MCB: Your users are going to have to fill in all 50 fields, and it feels like that’s going to take you a while. We have customers — I went back and checked, I think almost every single person you’ve interviewed on your podcast is a customer of ours. I don’t know if it’s 100%, but it’s definitely north of 95% out of the last 20 guests. Stratechery is a customer of yours, so there you go. MCB: Oh, really? Well, there you go. Thank you. One of my engineers adores Jira, so I get the opposite angle from what I asked about. MCB: That’s right. So look, it’s a challenge for sure, but at the same time, man, the value we’ve created, the business value, the number of customers that run on it, it’s ironic, we talk about the AI era and all these other things. Literally, no chips go out of any of the chip companies you love talking about, every single one of them, soup to nuts. So at what point did you realize that AI was going to impact you in a major way? Was there an “aha” moment or it’s just been in the air? Or is it a specific time you realized, “Look, this is going to completely change what we do?” MCB: Again, I’m one of these — I’ve realized I’ve become the old man in the room. We’ve done machine learning for a long time in lots of ways because of our online business model, so I’d say we’ve done AI for a long time. Obviously, LLMs are what people refer to nowadays by AI and agents and these words that have corrupted the entire thing, the meaning changes in technology when it means something else. The launch of various versions of ChatGPT were very instructive obviously, they were a moment for everybody. The optimism, and I would say we’re massive AI optimists, it is the best thing that’s happened to our business in 25 years. Why? Because people might look at you from the outside and say you’re still characterized as — even though your business expanded far beyond developers — “Oh, you have a lot of developers”, I’m skipping over the transition to the cloud just because we’re running out of time, but it’s an interesting story. You did announce you are finally ending the on-premises software, which I’m curious, it is a sentimental moment to come to that decision, but people might look at you from the outside and say, “Oh, there’s a company that’s going to have a problem with AI, AI is going to replace developers, it’s the decreased seats . What are they going to do?” MCB: There’s a few ways to take that. I’m trying to put it on a tee for you. I think I know what you want to say. MCB: There’s a few ways to look at it. Firstly, I think AI is a good example where people are very concrete about the negatives and the positives are upside. I think it’s a huge force multiplier personally for human creativity, problem solving, all sorts of things, it’s a massive positive for society. That doesn’t mean there aren’t any negatives, but the net effect is really high. And we spend a lot of time, you hear it in the media talking about the job loss, the efficiency gains, whichever way you want to put it, that’s the thing. Well, that’s because it’s really concrete in a spreadsheet, “I can do this process with half as many people”, “Wow, look at that, that’s great”, what’s never written in the spreadsheet is all the new processes that get created, all the new ways of doing things, the quality of the output is going to be twice as high. If software costs half as much to write, I can either do it with half as many people, but core competitive forces, I would argue, in the economy mean I will need the same number of people, I would just need to do a better job of making higher quality technology. So our view on AI overall is an accelerant, not a replacement to everything we do, and just the next era of technology change is really positive. We’ve loved technology, we love the cloud, we love all the tech changes we’ve been through, mobile. Look, us as a business, we are in the game of knowledge work. We solve human problems, workflows, business processes, this is what we do. These largely revolve around text, or if it’s video nowadays, that can be reduced to text in various ways. LLMs allow us to understand that text in a massively deeper way than we ever have been, and the problems we solve aren’t going away. 20 years time, there’ll be groups of people trying to solve some sort of problem as a team and working on a project, and so these things aren’t going to go. They’re going to need to talk to each other and collaborate of what work’s going on and how it’s working, so the textual aspect of it has been amazing. The features we’ve been able to ship, we never could have built five years ago, it was literally impossible, so the ability to solve customer problems is so much higher than it ever has been. Secondly, our software is incredibly valuable at the core of these workflows, but it’s also incredibly promiscuous. What I mean by that is we have always been very highly interlinked with everything else. If it’s a sales team, there are links to Salesforce and customer records, there are links to internal systems, there are links to maybe features that need to be built, there are links to some content and document. So any Jira, Confluence, or Loom , you don’t record a Loom unless you’re talking about something, you don’t have a Jira issue without pointing to all sorts of different resources, whether that’s a GitHub or Figma, whether it’s Salesforce or Workday. That gives us a really unique knowledge, which we’ve turned into the teamwork graph, that actually started pre-AI, so the irony is the Teamwork Graph is about 6 years old. Well, it started with Confluence. This is the whole thing where you look backwards, and to your point, if you had just been the Jira company, but because from the very beginning, you mentioned Confluence was different but it was adjacent and you had to build the links and stuff together, and as you build all these different tools, because everyone wants to be this point of integration. And I wanted you to tell me about Rovo and this idea of being able to search across all your documents. Who gets permission to do that? It’s someone that’s already there, and you made the critical decision to be there back in 2004 or whatever it was. MCB: That’s true. Certainly back in 2004, and then in I think 2019, the Teamwork Graph starts, which is trying to take all of those links and turn them into a graph. The connectivity, two things linked to this Figma thing, five things linked to this customer record — okay, cool, that means something, so we built this Graph. To be honest, it was a bit of a technology lark. We have a lot of these projects that are really cool and we’re like, “We’ll be able to use this somehow and it’s going to grown”, and now it’s a hundred billion objects and connections connecting all of the company’s knowledge. It becomes the organizational memory nowadays and context and all these things nobody knew in 2019 that’s what it was going to be, it just seemed we needed it for various process connections. That turns out to be because it’s got permissions and compliance and all of the enterprise stuff built in, which is incredibly difficult, the best resource to point AI at in various forms. You still have to be good at the AI parts to get the knowledge, the context for any area, so the Teamwork Graph is our data layer. It’s not only the best kind of enterprise search engine for your content from a 10 Blue Links kind of way of thinking. If you’re chatting through your content, you still need all your organizational knowledge. I actually obviously found your Article, I was like, “Hey, what has Ben Thompson written about us last year?”, and I asked Rovo in chat and it comes back to me with he wrote this, that and the other and pulls out some snippets. I’m like, “Tell me more, do you think we’ve hit that?”, I literally got a report written by Rovo on your report as to whether it had been accurate. “Go look at the last 10 years with deep research and web search and come back and tell me, was he right or wrong?”, and it gave me a really interesting analysis of whether you were right and wrong. It’s like most AI things, it’s like 90% correct, it’s pretty good. It solved a lot of the first problem and I would not have done that work otherwise. I would have read it quickly and so I wasn’t going to put an analyst on it internally to do this work, but I could send something to do work I never would’ve done. Who’s your competitor for this spot, for this Rovo position where you have all this context, you can actually search your company in a way that just wasn’t possible previously? MCB: Who are the competitors you say? Yeah, because everyone is claiming they’re in this spot, “We can be the central place that you go and we have visibility everywhere”, why is Atlassian the one that’s going to win that space? MCB: A few reasons why we will. I think we have a great chance to be a great player is maybe the easiest way to say it. I think everybody loves this absolute win position, we don’t believe in enterprise technology, you usually get these absolute wins, it’s not quite the same as in the consumer world. We have a lot of business processes and workflows, millions every day that run through us, those are human collaboration workflows, so they are cool. The auditing team hands off to the accounting team, hands off to the tax team, whatever it is, sales workflows, marketing workflows, and they span lots of our applications and many others. If you’re going to go and introduce agents, these autonomous AI-driven software programs, whatever you want to call an agent, you’re going to put them into existing processes to make those processes either more efficient, more accurate. When the human picks up a task, it’s got all the information they need because something’s gone out to find it, that is an incredibly powerful position, which is why we support our agents and everybody else’s. You can assign a Jira work item to a Cursor agent in terms of code, you can assign it to a Salesforce agent. If you have your agent technology choice, I don’t think you’re going to have one agent platform, I think you’re probably going to have multiples, there are going to be a handful of organizational knowledge graphs that are powerful enough to solve these problems across multiple tools, but we have access to all those tools. We already know the information to some level, and that becomes a very unique advantage. Do you see this as a way to expand even further how much of a company you cover? You started with developers, then you expand to adjacent teams, and you talk about it’s now just a fraction of your user base. Do you own entire companies or could you get there? It’s like, “Okay, we still have these teams over here that are not on Jira, but Rovo’s so good that we need to bring everyone in”? MCB: Look, again, it would be great. I think it is unrealistic, and we should say “Absolutely”, right? MCB: If [Salesforce CEO Marc] Benioff was here, he’d be like, “Absolutely, we’ll own the world”, we love him, that’s the way he is, I don’t think about it as owning a customer. Our mentality has always been — I always use the subway analogy versus we have some competitors, for example, that want to be the control tower, their whole thing is we’ll be the control tower, just give us control and we’ll go and control everybody else, we’ll move the planes around. I think in enterprise IT, that’s an unrealistic view. Every CIO has been sold this for decades, it doesn’t happen because the world changes too quickly. Our philosophy and our commitment to customers has always been we will be a great citizen on all sides, we will interact with all of the applications you need, the old ones and the new ones, and we will be a valuable point of exchange in your business workflows and processes, whether those are structured like in Jira, whether unstructured like in Loom or Talent or something else. The reason for that is you have lots of systems. We want to be a valuable station on your subway network, we don’t want to be at the end of one of the lines, we want to be one of the handful of hub stations that are about moving trains around, and that is the best way to get your knowledge moving in your organization, it’s the best way to deal with your processes. Therefore, we need to have amazing AI capabilities. We have a massive investment in R&D, we have thousands of people working on AI tooling at the moment, and we have a huge creation bent, which is one of the reasons I think — we’ve talked a bit about the data advantage we have, I think we have a huge design advantage, and I actually think design is one of the hardest parts of building great AI experiences because it’s real fundamental design for the first time. You had a great line, you did a podcast a couple of weeks ago that I’ll put a link to, but you mentioned basically, the customer should not need to understand the difference between deterministic and probabilistic in the context of design, that’s what you’re driving at here. MCB: They should not need to understand that, they should need to understand when outcomes, outputs may be wrong or may be creative. Again, you talk a lot about the fact that hallucination is the other side of creativity, right, you can’t have one without the other. Hallucinations are a miracle. We have computers making stuff up! MCB: Our job is to explain to a customer when that happens, so it’s like this might be something you want to do, and that requires a lot of design. We have a feature in Jira called Work Breakdown which is super popular, where I can take a Jira issue and say, “Make me a bunch of sub-issues, this task has to be broken into a set of steps”. I don’t believe in the magic button theory of AI, that I’ll just hit a button and it’ll do all the things, I believe deeply in the value from AI will come from human-AI collaboration in a loop. It’s me and the AI working back and forth. You talk about yourself and Daman quite a lot , and it’s you, Daman and ChatGPT working together, but it’s not like you ask one thing and it’s done. It’s an interaction, it’s a collaboration back and forth, and that’s going to happen everywhere. In Work Breakdown, what it does is it says, “Hey, based on these types of documents I’ve gone to find from your whole graph in Google Docs and Confluence, whatever, I think this piece breaks down into these, is that correct?”, and it goes, “No, actually, that one doesn’t make any difference, these two are really good, you forgot about this document”, “Cool, let me go do that for you again”, and come back and say, “Is it these?”, “That’s closer”, and then you’re like, “That’s good enough, it’s 90% of what I need”, and then I go add the two that I need myself. That is a huge productivity boost but it’s not magically correct, and it requires a lot of design to tell people, “These are not the answers, these are possible answers, help us refine them and get better at it so that you get the 90% upside and the 10% downside is managed”. Are all these people pursuing these full agents that act on their own, are they just totally misguided? MCB: No, because I think, well, agents will take — there’s a snake oil sales thing going on as there always is in any bubble, and the snake oil sales is not wrong, it’s just chronologically challenged. (laughing) That’s so good. MCB: Well, customers are struggling. When I talk to customers every day, they’re like, “Is everyone else using these things to just magically transform their business with this simple, it took them five minutes and it’s replaced entire armies of people?”, and I’m like, “No, nobody’s doing that”. What they’re actually doing is taking business processes that are really important to their business and saying, “Okay, can I make this step better? This is highly error-prone. It’s compliance in a large organization, how do I make this part of the process better?”, and we’re like, “Oh, we can totally do that”, and they will replace small bits of lots of processes so that in Ship of Theseus style, five years from now, the process will look radically different. Occasionally, they are replacing entire processes, but this is the 1% case, what they’re actually doing is they have whole machines that are running and they’re trying to fix this cog and fix that cog, and that’s super valuable for them. That’s not a downside, that’s really, really valuable. And often, it’s work they didn’t want to do, work that wasn’t getting done, it wasn’t done at a high quality, so we got to remember that, I say this quite a lot, people shouldn’t be afraid of AI taking their job, I fundamentally believe this, they should be afraid of someone who’s really good at AI taking their job. That’s actually what’s going to happen, is someone is going to come along, in a sales sense, they’re really good at using all these AI tools to give better customer outcomes or handle more customers at one time. Is this why you’re hiring so many young people? MCB: Yes, I guess so. Yes, they’re more AI-native, they come out understanding these tools and technologies. I find the biggest irony in universities is all these people who “cheat” their way through every assignment, I use cheat in quote marks, using ChatGPT to handle these assignments, and then they’re worried AI is going to take all these jobs. I’m like, “Wait, you literally took your own job of writing the assignment, but you’ve also trained yourself on how to use these tools to get the outcome required” — now one might argue the university degree should be different, but just like when Google came along and you could look up any fact, knowing facts became far less important than the ability to look it up. I still think AI, it doesn’t create anything, maybe slightly controversial, but I argue it synthesizes information, it’s really good at processing huge amounts of information, giving it back to you, changing its form, bringing it back. Humans are still the only source of fundamental knowledge creation. I point out one of the flaws in the one person billion dollar company argument, and this will happen but it’ll be an anomaly. That company doesn’t get created without that one person, so there’s not AI creating companies magically. It’s like can a company eternally buy back its stock? No, because at some point, someone is going to own the final share? MCB: That’s right and I think this is missed, right? This is where we say it’s about unlocking creativity and what we do for our customers is put Rovo and these amazing data capabilities that we have alongside all the enterprise compliance and data residency, and there’s a massive amount of making this work in the enterprise with trust and probity and security. It’s very difficult. And great design to say, “What do you hire us to do? How do you get these technology and business teams to work together? What workflows do you have in your projects and your service teams, and how can we make those workflows better with more data and make your teams more informed?” That will end up with us having more share of employees in a business that use our stuff every day. Awesome. You made two big acquisitions recently, the DX acquisition , I think, makes a ton of sense to me measuring engineering productivity, particularly in the area of AI. What actual ROI are we getting on this? MCB: And how much money am I spending? Because I’m spending suddenly a lot of money, right? This is not cheap at all, I have huge bills. Internally, we use Rovo Dev , we use Claude Code, we use GitHub Copilot, we use Cursor, we have them available to all. We have a huge R&D — again, I think we’re still number one on the NASDAQ for R&D spending as proportion of revenue. You can take that as a good thing in the AI era or a bad thing, everyone gets to choose their own view on that, but we’ve always been incredibly high on R&D spending since day one. The bills that we pay though are very high, so DX is simply saying, “Okay, cool, how do I measure what I’m getting for that? Should I pay twice as much money because these bills are worthwhile, or is there a lot of it that’s actually just it’s really fun and it’s not actually leading to productivity gains?”. This is going to be a hard problem because there’s a lot of money on the line at the moment that people are paying for these tools, which is not without value, but measuring exactly what the value is is really, really hard, and that team’s done a phenomenal job. And we now have an Atlassian office in Salt Lake City, Utah, where I already spend a lot of time. Totally by coincidence, but it’s really nice. So that purchase, love it, makes a ton of sense. In perfect alignment with you. How does The Browser Company fit in? MCB: A lot of ways. So I have believed for a long time that browsers are broken. We’ve built browsers for an era of software that we don’t live in today. And I don’t, in my browser, have a bunch of tabs that represent webpages, I don’t have that. I have a bunch of tasks, I have a bunch of applications, I have a bunch of documents, and the browser was fundamentally never built to do that. That’s what Arc, first product from The Browser Company — if you don’t use Arc every single day, you should be, it’ll increase your productivity instantly because it’s built for knowledge workers and the way that they have to actually work every day and how they manage all of these tabs and tasks and flows versus serving the New York Times or whatever. That is a browser built for knowledge workers, and there’s a lot more we can do in that era as software changes. Secondly, obviously AI has come along, and we now have chats and applications as a extra part of the browser experience, so I think we can change how enterprises use browsers, security being a big issue. I think AI in the browser is a really important thing, but I suspect it’s not in the basic way of just combining Chrome and ChatGPT, that’s not how it’s going to play out. I suspect it requires a massive amount of design, which The Browser Company is phenomenal at, and it requires changing how people use their day-to-day applications. From our point of view, and I’ve been an Arc fan since day one, [The Browser Company CEO] Josh [Miller] and I have known each other a long time, there’s a knowledge worker angle and there’s obviously a business angle to it in a huge way that our customers are knowledge workers. We can change the way they do their work in a meaningful way of productivity, that is exactly what we have been trying to do in a lot of different ways. The browser itself, being chromium-based, Edge being chromium-based, Chrome being chromium-based, the rendering of webpages is not the problem, it is the fundamental user experience of, “How do I take all of my SaaS applications, my agents, my chats, my tabs, my knowledge, and put it all together in ways that make my day quicker?” — that is what we are trying to do fundamentally at the start. The context that we have is incredibly important for that. And the browser has, if you think about it, my personal memory. We used to call it the browser history. Great, it shows what I’ve seen, it does not have my organizational memory, which we have a great example of in the Teamwork Graph. So if I can put these things together, I can make a much more productive browsing experience for customers fundamentally in that world. I think we have an amazing shot of doing that and of changing how knowledge workers use SaaS. We’re not trying to make a browser, as I’ve said, for my kids, we’re not trying to make a browser for my parents, we’re not trying to make a browser for shopping or for anything else. We’re trying to make a browser for people who spend all day living in Salesforce and Jira and Google Docs and Confluence and Figma and GitHub, and that is their life. The laptop warrior that sits in that experience, I believe we can use AI and design to make that a far better experience and build an amazing product. They’re well on the way to doing that, we can supercharge doing it. You look skeptical. No, I’m looking at the clock, I skipped over a huge section. Your whole shift to the cloud, all those sorts of things. However, there is one thing I wanted to get to: you are wearing an Atlassian Williams Racing hat , I am a big F1 fan, I was very excited about you doing this . How did that come about? How was the first year? Was this another hunch this is going to work out? I mean, Williams is looking like a pretty good bet. MCB: Yes, our world’s largest sports bet. Look, how did it come about? So how do I make a short answer? F1 is changing, I think, in a massive way. I know now being incredibly deep in the business of it, the fundamental change is that hardware is becoming less important and software is becoming more important, this is a trend that we are used to. JV, James Vowles , the Team Principal, was the first person that approached us a long while ago now to help them, and for a teeny, teeny sticker in the corner, to help them get more productive as a team. What people don’t realize about F1 is these are large organizations, right? There’s 1100 people that work for Atlassian Williams Racing. And Williams was really pared down and skinny, he was brought back in with new owners to actually rebuild the entire thing? MCB: Yes, they were in deep trouble. But in rebuilding it, he is a software engineer, software developer by trade, by history kind of thing. He’s a technically-minded person. He downloaded Jira himself in 2004 to install it, so he knows us quite well. So we were brought on for our ability to help them with their teamwork and their collaboration, they really needed a technical upgrade to a whole lot of their systems. Turns out they need us in almost every part of their business because the service workflow’s important. We’re now in the garage, we’re using tons of AI to try to make them better, so there’s a lot of things we can do to build to hopefully help them win, and it’s a mission you can fall in love with. Here is one of the most storied brands in Formula 1 that’s fallen on tough times, every sportsperson loves a recovery story. And I was sold early on the recovery story, I’m like, “Fuck it, let’s go help, let’s make this happen. Let’s get back to being a championship team”. So we fell in love with the mission, and JV is super compelling, he’s got a one-decade goal, and they’re very goal-driven, and we love that, but they needed a lot of help, so that’s what they asked us for help with is initially. The more we looked at it, the more we learned about Formula 1, yes, it’s becoming a software-driven sport. So as an example, Atlassian Williams, I believe have twice as many software developers as the next team on the grid. Because it’s cost-capped, you got to choose, “Do I hire a software developer or an aerodynamicist?” — it’s a very clear cost cap, you’re choosing where to put your resources. As virtualization and everything get better, it’s less, “How well can I draw a curve?” and, “How much can I help 1100 people work together, and how can we build great software”, which really is the core of the car, right? So that then comes to us, tiny sticker, probably a founder-ish moment where I’m like, “How much is the sticker on the top?”, and they didn’t have a sticker on the top and I’m like, well, “What would that get us?” So we ran the numbers on that and the reason is twofold. You talked about our GTM, our go-to-market transformation, we have an ability to build various things. Firstly, branding is obviously massive, top three teams get 10 times the branding as the bottom three teams. So if you’re going to make a sports bet, you pay for a long period of time with the bottom three team, you help make them a top three team, and your sport bet pays out really well just on a sheer TV time and etc — the number of staff, parents, and other things, have said to staff members, “Hey, that company you work for, it’s really great, I saw them on the TV on the weekend”, and the staff member will say, “Dude, I’ve worked there for 12 years, why do you suddenly know about it?”, “Oh, I saw them driving. Carlos [Sainz Jr.] is great”, or something. And he is! So obviously, there’s a huge marketing and branding angle that’s about their position being better. The really interesting part of what we’re doing there is we have customers all around the world, we have customers in 200-odd countries, and we can’t go and visit all of our biggest customers in a meaningful way. We certainly can’t take them to some of our best and most exciting customers, right? There are electric car companies that use our stuff that we’d love to take many customers to a factory, or rockets, or whoever, I can’t take many customers into some of your favorite chip companies and say, “Look how they use our stuff”, I can maybe get one or two customers a year into that customer and show them how they use our things. With Formula 1, what we’re building is a mobile EBC, so an executive briefing center. Formula 1 goes around the world. It goes to Melbourne, it goes to Singapore, it goes to Japan, it goes to England, it goes to various parts of Northern Europe, it goes to various parts of America and you’re like, “Hey, where are our customers?” — roughly distributed like that. It comes to town, we can invite a whole lot of customers into a great experience, we can tell them a lot about Atlassian software, we can also invite them into one of our best customers. They can sit in the garage, and I can tell them how our service collection is helping power the assets, that when that wing’s broken, it gets known here, and they start making a new one back in the factory in Oxford, and this one gets shipped around the world and another one will get moved. And, “Here, I can show you the asset management and the service that goes along with it, I can show you how the garage is getting more efficient because of us, I can show you how we’re helping them win races”. We don’t drive cars, we help them be more productive as a team and I can do that in an environment of it’s an exciting environment. They can drink a great latte or a champagne or whatever they want, and I can explain to them how we are transforming this business in a meaningful way with our tools no matter which way they want to look at it, which is the most powerful customer story that you can go and tell a couple-hundred customers a year in their city. We come to their city, right? I was in Montreal, I took a whole bunch of Canadian customers over the three days, they were like, “This changes my view of Atlassian”, and I’m like, “That’s exactly our goal”, that is at the enterprise end of enterprise sales though, right? But that’s the ironic thing, it’s as far away from where you started as you could be. MCB: Well, they didn’t get there. I met two Canadian banks we had in Montreal as an example, both of whom had been customers for over 20 years, they started spending $800 bucks or maybe $4800 as we moved our pricing to around five grand — now they spend a million, two million dollars a year, and they could be spending ten. We have the ability to give the massive business value across a far larger swath of their business. And I can say, “What do you use from our system of work today? What could you use? Let me show you how Williams uses that piece of the system of work”, which is just a very visceral and exciting customer example to show them how they’re winning. And it helps, again, culturally, super aligned. They’re an awesome group of people trying really hard to win in the most ridiculously competitive sport and the highs are highs, the lows are low. Any sporting fan, you’re well familiar with various different sports that we have in common, but this is technology built by a large business team that has to win a sport. That doesn’t happen anywhere else in the sporting world, I would claim. Giannis [Antetokounmpo] doesn’t make his own shoes and have a team of people making better shoes and a better basketball so he can win, that doesn’t happen in other sports. It’s all about the people on the floor in an NBA game as to who wins, and that’s great, don’t get me wrong, I love basketball. The work in Formula 1 is done by 1000 people back in Oxford. It’s a Constructor Championship . MCB: The constructor championship I do think should be more important, especially given the current exact week we’re in, which is an amazing week for Atlassian Williams Racing, second podium . You talk about that bet, I told JV at the start of the year, I thought that he’s like, “What do you think our five-year future is?”, and I said, “Look, I think, number one, we’ll get one podium this year, 2025; 2026, we’ll win a race; and by 2030, we will have won a championship, that is my OKRs [Objectives and Key Results]”, and he said, “Oh, wow, okay, yeah I think so”. It lines up, I know the team OKRs and other things. And we won two podiums this year, so I was wrong, and I think we have a great chance for 2026, and we are working hard to make the team better and the single-best customer example we have of every piece of software that we sell. Mike, I’d love to talk again. It was great talking to you again. And, hey, good luck. And I’m a Williams fan, so I’ll be cheering for you this weekend. MCB: Oh, yeah. Well, I’m not sure this weekend, but 2026, 2027- Okay. I’m kind of kissing up, I am dying for Max [Verstappen] to win is the honest truth. I need the McLarens to run into each other . But other than that, Williams is my second love. MCB: Do you think McLaren will issue team orders to switch them if Oscar is in second and Lando’s in fourth? Yes. And I don’t know what’s going to happen if that happens, and this will be fascinating. MCB: We will have to see. It’s going to be a huge week. But that’s what makes the sport exciting, right? The whole thing is amazing. Talk to you later. MCB: All right. Thanks, man. This Daily Update Interview is also available as a podcast. To receive it in your podcast player, visit Stratechery . The Daily Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly. Thanks for being a supporter, and have a great day!
Hardware
AI
Business
Sports
Career
Machine Learning
UI
Entertainment
Media
iOS
Mobile
Go
Sql
Java