The history and future of AI at Google, with Sundar Pichai

Transcript

Sander Pichai just passed a decade as CEO of Google. Alphabet is now not only one of the world's biggest tech companies, but a leader in the AI race, with plans to spend $175 billion in CapEx in 2026. Cheers. Cheers. Thanks for coming. Thanks for having me. A bit of history that people talk about a lot in the context of Google and AI is the fact that transformers were invented at Google, but then productized outside of Google, with mostly ChatGPT and that style of product. How do you reflect on that now? I think it's actually worth talking about. It's a bit misunderstood. Transformers were done in the context of a lot of TPUs. Transformers were all done to solve a specific product need to some extent, right? The team's thinking about how to make translation better. In the case of TPUs, how do you solve this? SpeechRack works, but you suddenly have to serve it to 2 billion people. We don't have enough chips for it. It's how do you solve inference for it? I hadn't known that. Transformers were specifically... It was from our research teams, right? [SPEAKER_01] But they were guided by solving product problems. Translation, yeah. And Transformers were immediately used. [SPEAKER_01] So BERT and MUM, people underestimate how much, because we measure search quality so religiously, some of the biggest jumps in search quality in that period where search went ahead of everyone else was because of BERT and MUM. We built Transformers and used it immediately in search to improve language understanding, understanding web pages, understanding your queries, kept building better models. We had also started productizing it internally in the form of teams building something called Lambda. Wow. Obviously, we weren't the first to ship that, but I think it's less to do with it just being research and us not applying it in a product direction. That I think is not the case. [SPEAKER_02] It's like, you did this research, you then saw massive ROI from using it the way you intended, [SPEAKER_02] and then you didn't invent all of the products that were invented with it, but that's to be expected. [SPEAKER_01] I would go a step further. [SPEAKER_01] We actually even conceived the product, which is ChatGPT. It was Lambda. If you would remember, there was an engineer inside who thought it was sentient, right? So think of it as an early version of ChatGPT you were speaking to internally. So we even had the product version of it in some alternate scenario. Google probably shipped that nine months later or something like that. Yes. In fact, at Google I/O in 2022, we launched something called AI Test Kitchen, and that was Lambda, but we had constrained it because internally, we didn't have an end-to-end version which was reinforcement learning trained, right? So the version I saw was a lot more toxic at a level we couldn't have possibly put it out at that time. Yeah. And also, as a company, which had this search quality bias, we had a higher bar, maybe, for what we thought was an acceptable product quality to go out, but it wasn't like we weren't figuring out how to get it out. I would also argue that even when OpenAI shipped, they did their deal with Microsoft probably a couple of months before. So you can look back and say it wasn't entirely fully obvious. I think they were lucky to also see it on the coding side with GitHub. I think there was a signal we were missing. On the coding side, probably you were seeing more of a sequential jump than on the language side. Yes. So the jumps between GPT-2 and 3 and later 4 were more pronounced if you were using it for coding. So you can point to things, but yeah. But I think to answer your original question, I think it was less about research to product than a bunch of other factors. I also remember talking to some of the people who worked on ChatGPT, and I think they launched it the week of Thanksgiving. It was a relatively limited launch. It wasn't presented as a big, prominent thing, like this is going to be an important part of our future. Clearly a surprise. I think it was an interesting test case. [SPEAKER_00] But the way I internalize these moments is if you're in consumer internet, you're going to have [SPEAKER_00] surprises. We were at Google then, Ilad and I. There was something called Google Video Search. [SPEAKER_00] YouTube came out, right? We acquired YouTube. Or think about if you were in Facebook, Instagram came out. Nobody sits and says, you don't look at those moments with that drama because Facebook just bought Instagram. Yeah, yeah, yeah. Right? But the way I've internalized consumer internet, people are going to be sitting and prototyping and throwing out millions of things. I'm not trying to diminish [SPEAKER_01] anything, but I'm just saying you're always going to have these moments. I don't think people wake up in [SPEAKER_01] a garage and ship a better iPhone. That's not going to happen, right? But that's not how consumer internet [SPEAKER_01] works. So you just have to be conscious of that and internalize that. As I think about the AI race in 2026, one thing that strikes me is Google has for so long had speed as the place it tries to differentiate. The original Google Search was really fast and famously displayed the search query time within the results, showing off. Then Gmail had fast search compared to competitors at the time, or Chrome compared to competitors at the time. And now, I mean, I use all of the AI services for different things, but Gemini on TPUs is just so fast. And I'm curious how much this is part of the explicit product strategy and how you think of it, or if it's much more nuanced than that. I've always internalized speed. Let's call it latency for this purpose, right? And as one of the distinguishing features of a great product. And also almost always reflects the technical underpinnings of the product having been done well. There's a different speed which matters too, which is the speed of shipping and iteration and release cycles. So both are important. But you know, you talk about latency. There are times when it's easy to say you want latency, but you're constantly adding capabilities. So the capability frontier is progressing. So there's some tension in how do you balance that. That's where it gets more complicated. But to give an example, like search, nuanced than that. I've always internalized speed. Let's call it latency for this purpose, right? And as one of the distinguishing features of a great product. And also almost always reflects the technical underpinnings of the product having been done well. There's a different speed which matters too, which is the speed of shipping and iteration and release cycles. So both are important. But you talk about latency. There are times it's easy to say you want latency, but you're constantly adding capabilities. So the capability frontier is progressing. So there's some sense of how do you balance that. So that's where it gets more complicated. But to give an example, search. I was speaking with the teams, right? They now have four sub teams with latency budgets in the milliseconds. You'll get 50% credit. So if you ship something which shaves off three milliseconds, you earn 1.5 milliseconds for your latency budget, and 1.5 milliseconds gets passed on to the user, right? And depending on what we think you're doing, some people may get a latency budget of 30 milliseconds or 10 milliseconds. You can use it. So you have rigorous reviews against that. But that's how much we think it matters. And for context, I guess humans pick it up in the low hundreds of milliseconds. Is that correct? In terms of where it actually impacts? [SPEAKER_02] That's right. Yeah, that's right. I think we've actually checked the dashboards in the metrics, and we've actually improved search latency by 30% in the last five years. But think about the functionality progression that's happened. This is why in Gemini, we deeply think about that Pareto frontier of making sure the capability to speed and the flash models are at 90% the capability of the pro models, but much faster, much more effective to serve and the vertical integration helps and so on. How do you think about the future of search, actually? Because a lot of people now are talking about chat as a new interface, obviously Gemini's incorporated, or search has incorporated Gemini or AI results in the context of Google. But a lot of people are now talking about agentic flows, and everybody's going to have a personal agent who, instead of typing in a query, it'll go and do something for you, instead of asking about trips, it'll go and plan a trip for you. What do you view as a future of search? Is it a distribution mechanism? Is it a future product? Is it one of n ways people are going to interact with the world? I feel like in search with every shift, you're able to do more with it. And we have to absorb those new capabilities and keep evolving the product frontier. If it's mobile, the product evolved pretty quickly. You're getting out of a New York subway, you're looking for web pages, you want to go somewhere, how do you find it? So you're constantly shifting that. People's expectations shift, and you're moving along. If I fast forward, a lot of what are just information seeking queries will be agentic in search. You will be completing tasks. You have many threads running. Will search exist in 10 years? Well, it keeps evolving. So search would be an agent manager, right, in which you're doing a lot of things. I think in some ways, I use agents today and you have a bunch of agents doing stuff. And I can see search doing versions of those things, and you're getting a bunch of stuff done. But I think that, your question is, if you think of search as a prompt that is not longer than one line, returning a bunch of different ranked results, as opposed to just telling you the right answer or something, I think your question is, does that product paradigm still exist? But today in AI mode in search, people do deep research queries, right? So that doesn't quite fit the definition of what you're saying, right? So people adapted to that, right? So I think people will do long running tasks, can be asynchronous. We all started, or life started as unicellular organisms, and now we have this complex life. And so the question is almost whether that former version or paradigm eventually goes away? And really, what was search becomes an agent, and your future interface is an agent, and the search box in 10 years or n years is no longer. The form factor of devices are going to change, IO is going to radically change. And so it's tough to think 10 years ahead. But we are fortunate to be in a moment where you can think a year ahead, and the curve is so steep, it's exciting to do that year ahead, right? Whereas in the past, you may need to sit and envision five years out. The models are going to be dramatically different in a year's time. And so riding the curve itself is exciting. And so I think it'll evolve, but it's an expansionary moment. I think what a lot of people underestimate in these moments is it feels so far from a zero-sum game to me, right? The value of what people are going to be able to do is also on some crazy curve, right? So once you view it that way, people would ask all these questions, right? When YouTube has done well, since TikTok and Instagram, you know, so you can give many examples. I think the more you view it as a zero-sum game, it looks difficult. It can become a zero-sum game if you're not innovating or the product is not evolving. But as long as you are at the cutting edge of doing those things, and we are doing both search and Gemini, right, they will overlap in certain ways. They will profoundly diverge in certain ways, right? So I think it's good to have both and embrace it. When we talk about search and where it's going and things like this, I'm reminded of the fact that a year ago, spring, summer 2025, sentiment was very negative on Google. The prevailing view was that search is cooked and we're going to have a really hard time. The core business model is under attack. Google was trading for 150-ish dollars a share. And now people have realized that's silly. Google has up and down the stack, whether it be applications or models or TPUs or whatever, as well as Waymo and YouTube and all the cool bets. What do you think investors, as a proxy for informed sentiment, misunderstood this time last year? Because clearly there was some big misunderstanding. [SPEAKER_00] You know, it was obviously very inward focused in that moment. So, prevailing view was that search is cooked and we're going to have a really hard time. The core business model is under attack. Google was trading for 150-ish dollars a share. And now people have realized that's silly. Google has up and down the stack, whether it be applications or models or TPUs or whatever, as well as Waymo and YouTube and all the cool bets. What do you think investors as a proxy for informed sentiment misunderstood this time last year? Because clearly there was some big misunderstanding. [SPEAKER_00] It was obviously very inward focused in that moment. To me, it was very clear in that moment, the Overton window shifted. I felt like the company was built for that moment. The vertical thing, it's not an accident. It was very intentful; we were in the seventh version of TPUs. I remember it might have been 2016 Google I.O. where we announced the TPUs and spoke about we are building AI data centers. This was 2016. The company was operating an AI first way. We had deeply internalized the shift. To me, we were behind in terms of frontier LLM models, but we had all the capabilities internally and we had to execute to meet the moment. The exciting part was when I look at it from full stack, we had the research teams, we had the infrastructure teams, we had all the platforms, we had been investing intentfully in many businesses. To me, I suddenly felt wow, we have this one common technology which can accelerate all those businesses. Search to YouTube, to cloud, to Waymo, all relies on progress. This was a very leveraged way to make progress. I understood it. To the earlier point of the discussion, I didn't view it as a zero-sum moment at all. I felt like everything is going to scale up 10x and there's going to be room for other people. You go back, Amazon has done well since Google came into the picture and Facebook. We underestimate the growth scenario of how all these things work. But we had to execute better as a company. That's what I meant by I was more focused on that. Was there something that demonstrated to the outside world that they got this? Was it Gemini 3 that changed people's minds? [SPEAKER_01] I think the real model probably where people saw it was maybe Gemini 2.5. [SPEAKER_01] And in getting to the frontier, particularly around multimodality, we made a bunch of this. Credit to the Google DeepMind teams. We paid a bit more of a fixed cost upfront, but we designed the Gemini models to be very multimodal from day one. Areas I think we started, the strength started showing. Nano Banana was an example of it. You were able to see it all together. But it's an amazingly dynamic frontier. I think there are two to three labs who are pushing each other pretty vigorously. At any given month, we feel like, oh great, we've done this well. Oh, there's a couple of things we're behind. But I think the picture will again be dynamic in a few months. The frontier is intense as you would expect it to be. It's interesting because when I talk to researchers not at Google or at the other labs, one of the things that they commonly bring up is that they feel the difference between two or three other labs and the Google team is that Google is not as AGI-pilled. In other words, there's less of a belief in AGI being right around the corner and the acceleration through it. Obviously, the folks at Google are thinking deeply about that. Do you think that's true? And do you think it impacts some notion of what the future actually looks like and therefore what people are building against? [SPEAKER_01] Look, we probably have scaled our capex from $30 billion to approximately $180 billion. Real money. You don't do it if you don't think about the curve a certain way. I view it as largely semantics maybe because we are a larger company with a lot of products that touches so many people at so many levels. Maybe the language of how we talk about it might be different. [SPEAKER_00] My earliest conversation. I think this notion that at Google we haven't understood what AGI is, Demis and team or Jeff Dean and team, at one point, Demis, Jeff, Ilya, Dario were all there. I like that retort. It's like, hello, have you been paying attention for the past 20 years? That doesn't make sense to me. I think some of it is if you're a younger company, or you were more a pure research lab, or you're headquartered in San Francisco, there are a lot of small attributes which can probably make a difference. But I don't think at a foundational level there is a difference in outlook on what the curve is or how we internalize the technology. Even within the company, there's a set of us living on the bleeding edge, firing agents, seeing what these things can do, seeing the agents pick up skills, do stuff, and also look back three months ago at what they could do now. We are living that exponential internally. Where I agree, you can point us at the history of Google. I think what a lot is getting at is a feeling. I saw a tweet saying, what you have to realize to explain what's currently going on in the valley is that every tech executive has severe AI psychosis right now and is spending a huge amount of time writing code and talking to AI and things like that. I thought that was a funny take and not without any truth to it. I'm curious what were your feelings on AGI moments along the way of the recent, or to what extent do you have AI psychosis these days? [SPEAKER_01] My first feeling of an AGI moment was 2012 when Jeff Dean demoed the earliest version of Google Brain. This is when the neural networks recognized a cat. That was 2012. I went with Larry to the DARPA challenge. It might have been [SPEAKER_01] is that every tech executive has severe AI psychosis right now and is spending a huge amount of time writing code and talking to AI and things like that. I thought that was a funny take and not without any truth to it. And I'm curious, what were your feelings about AGI moments along the way of the recent, or to what extent do you have AI psychosis these days? My first feeling about an AGI moment was 2012 when Jeff Dean demoed the earliest version of Google Brain. This is when the neural networks recognized a cat, right? So that was 2012. I went with Larry to the DARPA challenge. It might have been 2014, I think I need to be exact about when we went there, seeing the cars drive there. Demoing the earliest versions of the models having what we would call imagination. So there have been many moments like that. So it was obvious the technology is progressing. In terms of living now and having a visceral feel for it, I think the closest I would say is if you're coding and you give it a complex task and you never open the IDE and you're in some agent manager world and you see it do it and how powerful it is. So if you can call it feeling AGI, there are moments like that. [SPEAKER_01] Yes, yes. I did a little hobby project recently and after a while I was like, oh, I wonder what language it's using. But that was a detail that I needed to ask it about after everything was up and running. Yeah, you would just feel like magic. Yeah. So moments like that for sure. And the slope of the curve is what surprises you. Yeah. Right. And you're improving it on so many paradigms, it feels clear that there's going to be progress ahead. Right. When you talk about the visceral feel, I feel like one thing that's important to tech companies and every CEO thinks about this differently is how you stay connected to the product experience and everyday users because tech products are so abstract that it's easy to not just manage through reports from teams and slide decks and spreadsheets. And Tony Xu is talking about how he still works as a DoorDasher to stay very connected to that experience. We do at our company like a weekly all hands weaver segment of just walk the store where we click around in the dashboard together and we're tripping over like why is that modal there and that's confusing or whatever so we're collectively using the product. I'm curious how it works for you and how at Google you ensure that you're staying connected to the experience of using the products. Other than you use Gmail and everything every day. Oh yeah. You know, dogfooting like literally internal versions. I do block time to use it intensely. So kind of focus time to do it. And so that helps. Like even just two weeks ago, I was stretching in the gym and I had the phone with Gemini Live. And I'm like, I want to talk to it for the entire 30 minutes on one topic. So you do those things and some of it works, some of it is frustrating. But you learn a lot, right? So I force myself to use it in those power user mode ways and stay in touch that way. X helps because sometimes you get the raw feedback. [SPEAKER_01] Thank you for fixing the Google Calendar thing. That was so good. Well, there's a few more we have to fix. But thanks for flagging it. So yeah, X helps because you get the raw comments and I try to follow it directly. But I'll tell you what has helped. Internally, I would query in our internal version of Antigen. Hey, we launched this thing. What did people think about this? Tell me the worst five things people are talking about, the best five things people are talking about. And I type that. Now that brings it back. So has my life gotten easier? Yes. In the past, I would have to spend a lot more time trying to get a sense for it. Now an AI agent is helping me in that journey. So you can get, well, how much should I be spending firsthand to get that feel versus actually leveraging these tools? So even I'm going through a journey there, right? So I'm trying to adapt to this future. I guess there's, you mentioned A, that it's not zero sum. B, there's all these productivity gains people are seeing. And if you look at a lot of prior technology cycles, it took a while for the internet or for mobile or for SaaS to show up in actual GDP numbers, right? In the context of AI, we're seeing it from a data center buildup perspective, right? That's driving part of GDP growth. How do you think ahead in terms of three, four or five years, do you think the US economy is bigger because of AI? And if so, how much bigger? Look, for these returns to make sense, somewhere it has to, you know... How long was it before? I think it was maybe from Sequoia, someone wrote saying people are investing this much. Yeah, they're comparing the CapEx to the... Yeah. And this might have been two and a half years ago. And it was a talk saying it doesn't make sense because you would need to return at that level. You're probably 10x things. Yeah, yeah, yeah. Since that moment, I need to go look at the numbers again, right? So at some point it has to reconcile. To be very clear, we are supply constrained. We are seeing the demand across all the surface areas. I actually don't have any doubt that this is a massive market and outcome. So my question, and I think there's a lot of things that people misunderstand. So for example, people often talk about software engineering budgets and then what proportion of that is token versus salary. And to some extent, I think that market has been so demand constrained for great software engineers that suddenly adding supply can 10x that market, right? In other words, I think the market for software engineering and coding is dramatically bigger than anybody thinks. And it's the wrong metric to say token budget versus engineers. So I actually think it should grow a lot of things. I was just curious of your view of how much growth do we think is likely actually to come of this? I actually wasn't doubting at all CapEx versus outcomes or, you know? I see. Yeah. Look, I mean, going back to the internet and looking at GDP growth, you know, it doesn't quite capture what we all feel with the internet, right? And so maybe we would have had negative GDP growth without the internet. Consumer surplus. [SPEAKER_02] Yeah. So it's tough to look ahead. I do think there are natural dampening lot of things. Yeah. I was just curious of your view of how much growth do we think is likely actually to come of this? I actually wasn't doubting at all CapEx versus outcomes. I see. Yeah. Look, going back to the internet and looking at GDP growth, it doesn't quite capture what we all feel with the internet, right? And so maybe we would have had negative GDP growth without the internet. Consumer surplus. [SPEAKER_02] Yeah. So it's tough to look ahead. I do think there are natural dampening mechanisms in society at various levels. And the obvious ones being the compute build out is a different curve than the rate at which we can improve the models, right? So you're already dealing with a more constrained curve there. Then how do you diffuse it into society, right? We are doing this with Waymo, right? And you can make Waymo safer than human drivers, but you have to be careful at the pace at which we are rolling out, et cetera. So sometimes how do you diffuse it through society responsibly? There are constraints in all these layers, right? But I think the US economy is so much larger than it was 10 years ago. So to grow that, even at a half a percentage point higher, then that's a massive contribution. So I expect it to play out that way. Listening to Sundar is a powerful reminder of what it means to operate at true internet scale. When it comes to commerce, though, most businesses are forced to make critical decisions in a vacuum. They view the internet economy through only their own lens, using only the data that exists within their four walls. This is where Stripe comes in. Because we process $1.9 trillion a year in global payments, we have a panoramic view. When you build on Stripe, you're plugging in to that network intelligence. From verifying a user's identity, to stopping fraud before it starts, to streamlining how businesses connect and knowing what payment methods work best locally, we've built a system that puts 1.6% of global GDP to work to protect and grow your revenue. So if you want to use the power of Stripe's network intelligence, come see what you can build here. You referenced the supply constraints. And I think that's a really interesting defining aspect of 2026, where you said 150 billion in CapEx, 180? We have said it'll be between 175 and 185. Okay. So 180-ish billion of CapEx. And what's interesting to me is that Google could not spend $400 billion in CapEx if it wanted to, because the memory isn't there, and the power isn't there, and all these components. So can you just tick through? We can find a number of electricians we would need. [SPEAKER_00] Exactly. So I'd love to hear just your overview of the various bottlenecks. [SPEAKER_00] Look, at some level, you have to work back to actual wafer capacity or something like that. So there are deeper ground troops. I think wafer starts is a fundamental constraint. I think power and energy are more solvable. Permitting and actually working through a regulatory environment might be a constraint. So the pace at which you can do things. Even though there's lots of land in pro-growth Texas or Nevada or Montana, just maybe not enough? [SPEAKER_00] I think we're making tremendous progress. I think for the US, I think it's a particularly important thing. You're in awe of the pace in China, how fast they can build things. So I really think we need to learn to build things much faster. You almost have to shift your mentality to think about what would it take to do things 10x faster in the physical world? Construct 10x faster. But I would worry about that as a constraint. I think there could be growing resistance. So it's not as simple as a few people deciding you want to build faster. The data center moratorium and stuff. Yeah. Yeah. So I would say wafer starts, the ability to permit and do things. And I do think there's a lot of good work being done from the government on. I think people realize you need to do these things better. Then comes critical competence in the supply chain. Memory is a good one. We are constraining those things in the short term. Everyone will respond to it. But I think all of us running companies, regardless of how AGI-pilled you are, then comes this error bands of how bullish can you be? What's the margins you can afford? Because there are extraneous factors which can go wrong in the world, which are outside of control. So everyone is making those adjustments. Those are all constraints, right? And constraints. So I think that's where I see the constraints. Is memory the biggest component that you think about? Memory is definitely one of the most critical components now. Yes. And you said in the short term, do you think just people ramp up supply and so high prices will take care of us? There is no way that the leading memory companies are going to dramatically improve their capacity. So you have those constraints in the short term, but they get more relaxed as you go out. Yes. But I do expect all of this to constraint. By the way, I think it'll push a lot of innovations on, we will make these things 30x more efficient. Yes. All that is happening simultaneously as well. Does that enforce an oligopoly market? So if you actually look on the model side, because if you look at a lot of the views of models and how they're going to improve, a lot of it is going to be both self-improvement. So the models will start writing more and more pieces of themselves, do more data labeling for themselves, et cetera. So it's a musical chairs game of who has compute right now, basically? [SPEAKER_02] Exactly. Who has compute right now and how much can you actually scale relative to overall industry capacity? And if everybody is roughly pro rata up to some number, you've effectively put a ceiling on how much farther ahead somebody can pull versus everybody else. Do you think that's a correct statement or an incorrect statement? I think it's a reasonable framework to think about it that way. But there are things which are, I'm coming here as we just shipped Gemma for, right? And it's a really good open source model. The Chinese models are very good. Yeah, but anything outside of China, it's a very good open source model. The frontier to Gemma 4 is both huge and not so huge in terms of time. Gemma 4 is based on Gemini 3 architecture, right? It's a very weird thing, right? You're talking about a set of weights which can fit on a USB stick. Yeah, yeah. It's amazing. Right? So it's a really crazy thing. It's not like a SpaceX rocket. I guess, I'm always shocked that you run a data center for months and months and months and then your output is a file. Yeah. Literally. But anything outside of China, it's a very good open source model. The frontier to Gemma 4 is both huge and not so huge in terms of time. Gemma 4 is based on Gemini 3 architecture, right? It's a very weird thing, right? You're talking about a set of weights which can fit on a USB stick. Yeah, yeah. It's amazing. Right? So it's a really crazy thing. It's not like a SpaceX rocket. Yeah, I guess I'm always shocked that you run a data center for months and months and months and then your output is a file. Yeah. Literally, it's like having a Word doc or something and that's your model. It's amazing. So there are these unique attributes about this. So which makes me challenge those frameworks and say, how should we think about this? But I think it's a reasonable, at least on the inference side, what you're saying is a very reasonable way to think about it. But I do think everyone is trying to figure out how to blow through the capitalist incentive to break through these constraints. Yeah. It's immense. But as you say, there's only so much memory in the world. So no capitalist incentive will really solve 26 or 27 memory supply. That may be the era where you see more divergence. Yeah. So, and remember that has to balance with wafer capacity increasing, you being able to permit those data centers. So this constraint may be less severe than it appears. Right? So you have to envision the total square set of all the things that you need and then think it through, right? Are you worrying capital? Yes. Yes. But what's interesting to me is that plausibly people would invest beyond the current capex, but we're now just running against 26 and 27 real world constraints. It's a little about the strait of Hormuz. You can have whatever price of oil you want. Ultimately, if you take 20 million barrels a day out of the system, you need to destroy 20 million barrels a day of demand. And it's kind of similar with memory where ultimately some people have to not get the memory they want. But there are other constraints, right? Like which, take security as a constraint. And these models are definitely going to break pretty much all software out there. Maybe already we don't know. Yeah. Yeah. Yeah. I'm not sure if you can hear and speak. Do you really think all software there? Because SSH, people have been trying to break for a long time. Do you think just- [SPEAKER_01] No, I'm not talking about just think. Just regular software. [SPEAKER_01] Software, large platforms, right? How many zero days? Yeah. Yeah. You know, so there are constraints here in the system, right? You just can't wish away. Right. Somebody was telling me that the black market price of zero days is dropping because the supply is growing due to AI, which I thought was a really interesting market metric. Yeah. Not at all surprised, right? And not at all surprised. So, but when, how does it practically diffuse through society? What are the implications of it? Yes. Right? And so there are parallels, I think. So I think there could be hidden constraints. Yes. And there could be shocks to the system, if you will. But having said that, I genuinely think there's a lot of upside ahead. Some of the constraints maybe are helpful. Yes. Right? I think constraint inspires creativity. For the compaction cycle where we get more efficient, for the forces maybe important conversations to be had, which otherwise wouldn't happen. Yeah. Right? I think on my security point alone, I thought about, we are going to need more coordination. Yes. Which is not happening today. There will be a moment of, it could be a sharp moment, right? And so all those things, I don't think you can wish them away. Yes. Yes. Right. Yeah. Actually, related to that, Google does have an amazing portfolio of things that's both built and bought into. From an ownership perspective, you own a reasonable amount of SpaceX, I think. I don't know the exact amount, but I think it was 10-ish percent way back when. Anthropic, 10-ish percent, the majority of Waymo, which is an amazing thing. And then internally, obviously, there's this enormous swath of amazing technology that's been developed. We talked about AI and transformers. There's TPUs. Obviously, Waymo was another one of these things. There's Quantum. You just released a very interesting result there. Are there other hidden gems that people should know about or that are especially interesting or that may have very big impact in the future? Or you're peaking people maybe underestimates. Yeah. Yeah. Look, we're constantly trying to take these long-term projects, which, when you first announce them slightly marginally, looks ridiculous. We're in the earliest stages of thinking about data centers in space, right? But your earlier discussion around constraint inspires creativity. But if you take a 20-year outlook, right, where are you going to put most of these data centers? Really hard problems to solve. But those are examples of projects we think about today, which are Waymo in 2010. Quantum itself is one of these projects. We are in a deeply committed way making progress there. And I'm excited about it. Where do you think Quantum will have the biggest impact? Because mainly people talk about molecular modeling. They talk about cryptography. There's quantum proof cryptography that people have been developing over time. On the molecular modeling side, it actually looks like the deep learning models tend to be very good at that in certain circumstances. I mean, you all pioneered that with AlphaFold. Do you think Quantum will actually matter? And if so, where do you think it'll have the biggest impact? Look at abstract level, to me, it feels like to simulate nature more and more. Given it's inherently quantum, you would need quantum systems to better simulate it. We may get there with classical computing techniques in a surprising way or get at it with enough compression and abstraction. It may work. But I fundamentally felt like Quantum would have an edge there. And I don't know, we still don't understand the Haber process for fertilizer. There are many complex processes. I mean, it's probably your background going back to what you did in college. [SPEAKER_01] that with AlphaFold. Do you think Quantum will actually matter? And if so, where do you think it'll have the biggest impact? At an abstract level, to me, it feels like to simulate nature more and more. Given it's inherently quantum, you would need quantum systems to better simulate it. We may get there with classical computing techniques in a surprising way or get at it with enough compression and abstraction. It may work. But I fundamentally felt like Quantum would have an edge there. And we still don't understand the Haber process for fertilizer. There are many complex processes. It's probably your background going back to what you did in college. So my instinct tells me there'll be simulating weather, simulating reality, all that. I think Quantum will have an advantage. I think the way the history of technology works is you get something to a scale where it works and then you use it and people's creativity on top finds the applications. I always give this example of mobile phones plus GPS enabled Uber. There's nobody who was working on phones who would predict that as an outcome of this platform shift. So I'm confident Quantum will have many, many applications if you can actually make it work. That's how I think about it. So we're talking about your favorite projects at Google further afield. The Gemini team is deeply thinking through robotics. Robotics is an area where we were too early as a company before. It turned out AI was the missing ingredient for a lot of ideas, maybe 15 years ago or 10 years ago. But the Gemini robotics models focus on spatial reasoning, et cetera. So we definitely have state of the art models there. And we are partnering, in an ironic way, with Boston Dynamics and Agile and a few of the companies and in a determined way making progress. And there are extraordinary startups out there as well. But we are investing in quantum data centers in space, drone delivery with Wing. I think we are scaling up Wing where in some reasonable time period, 40 million Americans will have access to a Wing delivery service. I'm not talking years out. But these are all methodical compounding when you take these long term projects. We are committed. Isomorphic is very exciting. Think about being focused on these models in a targeted way to improving all the possible steps in drug discovery. And even though you have long pulls, like phase three trials, getting there with a much higher probability of success. I think it's definitely the smartest approach I've seen in terms of the different bio models and really thinking about the broader swath beyond just the molecular design, which is where most of them are stuck. It seems very smart. Can I ask, I'm curious, how capital allocation actually works at Google? And what I mean by that is the idea of good capital allocation is about internalizing that the opportunity cost for capital and putting the cash that a business generates towards its highest and best use. And in the toy example in a business school book, you know, maybe you're Boeing and we can either have this cash that our business generates and we can either go bid on the next defense contract and we'll invest this much in R&D dollars and we model this much revenue from the contract, or we go develop a clean sheet commercial airliner and we'll put in this money and we model this amount. It's like a 16% IRR versus a 19% IRR. Okay, I prefer the 19%. In Google's case, the projects are extremely heterogeneous. It's like, okay, we can give the YouTube team more funding so they can improve the recommender algorithm and therefore time on site increases and so does monetization. Or we can give the Waymo team more funding so that they can actually get to market faster or scale up faster. Or we can invest in this new AI approach that might pay off in five years time. And so I'm curious, if you are trying to put capital towards the highest and best use and you're ultimately comparing, how do you compare initiatives that are so different in nature and so different in payoff curve shape? That's the most John question ever. I need to know the answer. You need to throw an ROIC and then it's like— It's a good question. Look, I feel it today more than ever, ironically, because of TPU allocation. [SPEAKER_00] That's interesting. Yeah. So in some ways I feel it even more urgently. Needs TPUs, right? [SPEAKER_00] That's interesting. Yeah. [SPEAKER_00] So computers made the question, ironically, much more front of mind. [SPEAKER_00] By the way, of all the things I do, I'm really looking forward to how AI as a companion at least gives inputs to this task. And I think once we can actually get all the data connected and flowing through, I think the models are already capable. It's more getting all the data unlocked. I think will be helpful. So I feel it there. Historically, I think at Google, one of the advantages we have had is sometimes we make these decisions very early in the cycle. So it's almost like going back to that route, there's a deep technology orientation. And we actually think about the question you were asking a lot about what are those longer term things. And so I think thinking at that stage, it's easier because your initial funding amounts can be smaller. But then you stay committed for the long term, and you're making sure you're making progress in a deep way. So as long as you're seeing that underlying technology—like take quantum, for example—how do we judge it? We're judging the underlying, the milestones around what logical qubit error-corrected, large stable logical qubit threshold by when you're going to get to. Is the team able to do that? Right? So I think you assess it that way. So one of the ways we have thought about it and we've been disciplined about—or at least to me matters a lot—is to make those early technology bets in a deep way. And that's helped. But on a constant basis, look, I've always viewed it as you have to assess the long term value of these things. So it's almost like in some intuitive way, you're thinking about the option value and the time of something five to ten years down the line. And you assume crazy growth, right? And think through whether those decisions make sense. So the TPU investments have been great that way. And we've steadily about it and we've been disciplined about, or at least to me matters a lot is to make those early technology bets in a deep way. And so that's helped. But on a constant basis, look, I've always viewed it as you have to assess the long term value of these things, right? So it's almost like in some intuitive way, you're thinking about the option value and the time of something five to 10 years down the line. And you assume crazy growth, right? And think through whether those decisions make sense. So the TPU investments have been great that way, right? And we've steadily invested in that. Waymo was a great example where I think we increased our investment two to three years ago when the rest of the world got pessimistic on it. When others, some people were backing off. It's very magical. It's such a magical experience. I take Waymo now every day to work when I can. And it's magic. [SPEAKER_00] So I think Waymo's a good example of this. I have this question, which is Google does cut projects. And there are various things you've tried where you said, you know, we're actually not going to fund this part of X all the way, or we're not going to retire this product. It's not working. But Waymo, despite the fact that it was a long road from a compelling demo to commercial service in market, you guys didn't lose the faith. And so what was it that you were seeing? Is that a qualitative decision or a quantitative decision? How do you decide that we're going to cut Loon, but keep Waymo? I think it's to do with some kind of quantified assessment. You look at the Waymo driver, that's underlying technology, which is how does the software drive the car and the progress in terms of safety and reliability. So it's a long running task, how safe and how well you do it. And you follow that curve and you predict or you set goals where you want to be and how you perform against those curves. I think the team has been phenomenal. There have been phases where it didn't progress, but those are the times you need to have confidence in the quality of the team to break through those phases. But I think the more you're able to evaluate things at that deeper technology level, I think you tend to make those decisions better. Or at least that's how I have tried to do it. One argument I've heard is that a lot of the huge gains that have been seen recently is because it used to be hand mapped heuristics of how do you deal with edge cases of driving or something happens and how do you respond. And a subset of those were almost hand drawn out for the cars to follow. So a narrow set of things that it could do. And then the breakthrough was moving to end-to-end deep learning a couple years ago as this big transformer wave was happening in general. Do you think if Waymo had been started five years ago, it'd be at the same place as it is relative to having been started 15 plus years ago? Just given that that's the breakthrough that's propelled it forward? Look, I think we spoke earlier about robotics. You can think about Waymo as a robot, right? I think people who are starting robotics in the last three years, by definition, would be making faster progress, maybe. But I think Waymo is such an integrated system. There are aspects of it, not quite like, you know, something complex like TSMC or SpaceX launching things, you are talking about system integration in these things in a very complex way. I think Waymo has hidden aspects of that, which the time of how you do it, the craft of it matters. But having said that, I do think the end-to-end approaches are going to be an accelerant in this series. Because just having a team arguably was a huge benefit to Alphabet and Google, right? I mean, just the fact that you kept investing in it, and then it hit a moment in time where this technology liftoff was more than worth it, and was very smart and forward thinking. I just think it's interesting to ask how that applies to other domains. Because to your point on robotics, it seems like robotics will potentially have a different history where you can move very quickly now. Do you folks think about re-internalizing hardware again? Or is it largely going to be a partner-driven model to bringing this stuff to the world? I think we'd keep a very open mind. My lesson from Waymo and on the AI side with TPUs, I think to really push the curve well, particularly in areas where you have safety, regulatory, everything. You want the first-hand experience of the product feedback cycle. So I think having first-party hardware will end up being very important, is how I would say right at this stage. Makes sense. So I have two more capital allocation questions. Can you make the case that Google has historically been under-levered, where Google has historically carried a strong net cash position? And given that both Google has more ideas than it knows what to do with, it's just brimming with good ideas, and the core business grows very durably, and I think Google clearly has a very good understanding of that core business, and it has grown at a higher rate than Google's cost of capital. As you look back on it, should Google have been more leaned in and said, okay, we will be willing to have a leveraged position that's slightly more aggressive than strongly net cash, and we will put that towards new initiatives, or just buy more of this core Google business for Google shareholders, or do more minority investing, which Google seems to have been best in class at? It's a great question. For example, if Waymo had reached this point earlier, I think I would have invested the capital earlier. So to some extent, I think you were judging it by, like you want to be good stewards of capital. So to the extent you're bullish on ROIC, you want to invest every last dollar you can there. But to the extent you have access where you don't think, I mean, this is why we've invested in other companies too, right? Even if not then, but we've always thought about it with the lens of being good stewards of it. We felt our investment in Stripe was being a good steward of our capital. SpaceX, right? You know, SpaceX and Anthropic and so on. So I think now with the AI shift, there are more opportunities on which we can deploy capital in a good way. And so we're doing that. Yes, yes. But I think we always had that mindset. But I would have been glad to invest more capital in Waymo earlier. But we weren't at the level of maturity needed to do that. There was a point in Waymo from a safety standpoint, you know, we did approach Waymo safety first. And you just, it wasn't the right thing to do. So you feel like you cannot point to projects where they would have [SPEAKER_02] it. We felt our investment in Stripe was being a good steward of our capital. [SPEAKER_02] SpaceX, right? SpaceX and Anthropic and so on. So I think now with the AI shift, there are more opportunities on which we can deploy capital in a good way. And so we're doing that. [SPEAKER_02] Yes, yes. But I think we always had that mindset. But I would have been glad to invest more capital in Waymo earlier. But we weren't at the level of maturity needed to do that. There was a point in Waymo from a safety standpoint, we did approach Waymo safety first. And it wasn't the right thing to do. So you feel like you cannot point to projects where they would have gone faster had they gotten more capital sooner. They just needed a natural ramp. [SPEAKER_02] I wouldn't say that. But I think in general, at least, we might have gotten the decision wrong, but our approach at least was to say, if you got excited about something and had the conviction, yes, we were willing to come in with the capital to see through. [SPEAKER_02] And my other capital allocation question was, historically at tech companies, the large majority of the R&D expense was the people walking around the building. And headcount was managed through a very tightly controlled process. And indeed, as you thought about [SPEAKER_01] allocating R&D effort, it was really allocating highly paid people to go work on the challenge. And the tech costs were, unless you were doing something very computationally expensive, which obviously Google did, you know, Google Books or something. But broadly speaking, the tech was an afterthought compared to the cost of the people. We're now going to a world where, as you say, that's not the case with TPUs and how you allocate that. Just at a very concrete budgeting level, how does that work inside of Google? Like, do you have an overall TPU budget for the company? And then when you are giving a project resourcing, previously you gave a certain headcount budget, and now you give it a headcount and a TPU budget. Are they the same budget? Just how does that work when you're doing a quarterly review or an annual review? [SPEAKER_01] Look, we've always had a compute budget. Now, we've always had a compute budget, right? Even classic compute. I would say with ML, and we use both TPUs and GPUs extensively. But ML compute planning is, we are super thoughtful about headcount planning too. But we've always had to plan that. And ML compute, we've gone through phases where they've been easy. And then there have been phases where we've been constrained as a company. But now it is really acutely constrained, right? So you spend a lot more time. I at least spend a dedicated hour a week thinking about that question at a pretty granular level. So I will know by projects and by teams, the compute units they are using, right? And I have that information, and I'm looking at it and assessing it. And in some ways, it's a really important thing to be doing right now, I feel. How do you learn? So the scarce resource is compute in a lot of cases. And so you're ensuring that Google's precious compute resources are being spent on the most worthwhile initiatives. [SPEAKER_01] That's right. Yeah. How do you think about that in the context of GCP and Google Cloud? Because there you're actually allocating the compute to others instead of for your own purposes. And given the constraints in the system, how do you think through that differential allocation? [SPEAKER_01] Look, plan ahead, right? So when we do the forward planning, the cloud team is forward planning and they're putting a plan in place. And so you're funding that and you're doing that for our internal needs. You forward plan. And as part of that, you're also signing long-term commitments to customers. Anything we commit to a customer is sacrosanct, right? So these are contractual commitments. So you solve a lot of it with planning. And so there are, when you plan, we're all in a constrained world. So I think the cloud team would say they don't have the compute they want. But you solve it with planning ahead. [SPEAKER_01] Speaking of Google Cloud, I have my product request that I've been saving up for this section that I know you're looking forward to. You could have posted it on that. [SPEAKER_01] Exactly. Yeah, yeah, yeah. But I'll say one thing that works really well is the GCP MCP is awesome, where your AI can just interact programmatically with Google Cloud. And I guess you guys have exposed almost everything except the core, permissioning stuff. And I feel that, in a way, part of the curse of Google Cloud has been there is so much functionality there that I'm sure you occasionally hear from people that it was a little hard to navigate. You log in, you have to create an organization, a project, and whatever, and find the right services. And now, all that doesn't matter. And so you just say, hey, go add this Google Cloud functionality. And so that is something that actually, it feels like Google Cloud is really benefiting from the fact that it is so broad and there's so much functionality there. I mean, we have a little bit of this problem with Stripe, where as we add more functionality to it, just the right way to navigate this big product surface area is an AI that's read all the API docs for you. So that's working really well. [SPEAKER_01] I mean, the promise of AI being this orchestration layer, for anything you think about, to my earlier question, even internally within the enterprise as a CEO, it's not that you don't have all the data, but how do you get it in one place and see it? In the past that would have meant one more big ERP-ish project to go connect all the data sources. Again, AI being this orchestration layer in a way that makes sense for the end user, I think it's been delightful to see. [SPEAKER_01] So the bigger the product surface area, the more that benefits. And again, we've seen that to some extent with Stripe, but I feel like with GCP, it must be a massive effect. [SPEAKER_02] I think we could do a lot better. So, but you're right. It's an immense opportunity, I think. Yeah. [SPEAKER_01] I've been really happy with it. Okay. And then that gets to my products. Did you bring product suggestions for us? [SPEAKER_01] No, you go first. Yeah. I wanted to, but. What's interesting to me about open claw and the product market fit of things like that is they're allowing stateful AI for consumers. And if you want to say, you know, the classic, you know, round up the daily news that I'm interested in and send to me each morning, [SPEAKER_02] I think we could do a lot better. So, but you're right. It's an immense opportunity, I think. Yeah. I've been really happy with it. Okay. And then that gets to my products. Did you bring product suggestions for us? No, you go first. Yeah. I wanted to, but. What's interesting to me about open claw and the product market fit of things like that is they're allowing stateful AI for consumers. And if you want to say, the classic, round up the daily news that I'm interested in and send it to me each morning, or just something that involves persistence, that none of the popular or mainstream AI apps allow persistence. Is that common? I think that actually, look, I think you want to give users capability where you have persistent long running tasks. Yes. In a reliable, secure way. You have to think through things like identity, access, et cetera. But I think that's the future. That's the agentic future. [SPEAKER_00] Mm-hmm. [SPEAKER_00] And bringing that for consumers is an exciting frontier we are looking at. [SPEAKER_00] Yeah. This is one of mine too. This is Dreamer, which was the former CTO of Stripes company that just got bought by Meta. I think that a very good version of this. [SPEAKER_00] Mm-hmm. It's a very early view of. [SPEAKER_02] Yeah. They were making custom software, including persistence, but also, you could spec out. You could make your own little app. Exactly. Yeah. Yeah. And they made that very easy to use. But I feel like when people have this experience, there's a surprise and delight moment. And it's just interesting to me that. [SPEAKER_00] Look, I think effectively the consumer interfaces are going to have full coding models underneath, right? And the right harnesses and the right scales. [SPEAKER_00] Yeah. Yeah. [SPEAKER_00] And the ability to persist and run somewhere securely in the cloud, locally and in the cloud. So all those primitives are coming together. And so what developers are like today, I feel like there's 1% of the world, maybe not 1%, 0.1% of the world who's kind of living this future. Mm-hmm. Right. Mm-hmm. They are building stuff for themselves. But bringing that to mass adoption Yes. Is a very exciting frontier, I think. [SPEAKER_01] Okay. My other product suggestion is, sorry, you have to endure this part of the interview. Awesome. It's the right of passing. [SPEAKER_01] Exactly. My other product idea is, for some reason, I don't know if this is your lived experience, but certainly my lived experience, that searching Google Docs is so much harder than, say, search in Gmail. And obviously, they're both equally good search engines. But I think what's going on is keyword search works reasonably well for email, because you can probably remember a unique set of keywords for that email. Whereas what always happens, at least to me, is I want to go back and look at the 2026 budget. It turns out if I search Google Slides for 2026 budget, neither of those words is particularly unique in the context of words that exist in PowerPoints at Stripe. And so, I can never find the exact right one. And I'm curious, does Sundar Pichai also have this problem? Yeah. Somehow, I haven't felt it as acutely as you're describing it. But when you describe it, it resonates well with my experience. I'm literally playing through the person to whom I'm going to play this segment of the conversation. I know exactly who I'm going to go talk to. The people are working on it. I think we can make it a lot better. I think, look, the AI integration into these services, including Google Docs, I think you will see sharp improvements in the coming months ahead. [SPEAKER_00] I think we all did the first versions of it where you just put it in somewhere. But I think, over time, what all can you keep in context? What can you cache? And what can you really bring to bear? I think we can make a lot of progress on. So, I think we can do a lot better. [SPEAKER_00] Okay, great. We have a good extra putting up with this. A lot of companies that I'm involved with, even ones that were started reasonably recently, have had to dramatically shift their workflows relative to product development, engineering practices, who they even think of should be on the design team and the capabilities of that. Are you revisiting all that at Google? Are you rethinking it? Has there been big shifts in workflow or other aspects? The way I would say it is, you can think of it as concentric circles. There are some groups within Google who are shifting more profoundly. And so, for me, a big task was, how do you diffuse that to more and more groups, particularly in 2026? Some of it, we couldn't do it early because it breaks so often that, you know, it's almost like you see this promising new world, but it's kind of semi-broken. But this year, I feel like the curve is shifting pretty dramatically. So, I can see groups, particularly, I would say, GDM and some of the SWE groups really change their workflows, right? And, you know, they are using, we call this for some strange reason, we have a different name internally than externally of the same product, but it's JetSki internally, which is anti-gravity. And you're living on it, you're living in an agent manager world, you have workflows, and you're working in this new way, right? But just last week, we rolled it out to the search team, right? So, we're constantly pushing that. You know, in a large organization, I think change management is a hard aspect of this technology diffusing, which may be easy for a small company, right? You can quickly switch over. Can I lay out a few problems I see when it comes to actual diffusion of AI in industry? [SPEAKER_00] And I'm curious how and when you think we'll solve them? Because I see it with a big intelligence overhang. Like, the AI's are now amazing in terms of what they can do in the abstract. And if you look at how AI native a company is, or just how much it uses that intelligence, there'll probably be a shortfall. And the problems that I see are something like, one, it actually takes a while to get good as an engineer at prompting your AI well. And you can prompt an AI better or worse to write code. Then there's a lot of, say, Stripe-specific prompting in our case to know which tools to use. And so there's the general being good at prompting, and then there's the Stripe being good at prompting. And then, of course, you have the fact that it's hard to share an AI-generated code base because you have a blast radius, and you're just changing so much, and the turnover of the code At how AI native a company is, or how much it uses that intelligence, there'll probably be a shortfall. And the problems that I see are something like, one, it actually takes a while to get good as an engineer at prompting your AI well. And you can prompt an AI better or worse to write code. Then there's a lot of, say, Stripe-specific prompting in our case to know which tools to use. And so there's the general being good at prompting, and then there's the Stripe being good at prompting. And then, of course, you have the fact that it's hard to share an AI-generated code base because you have a blast radius, and you're just changing so much, and the turnover of the code is high enough, or maybe you're rewriting it several times before you ship, that it's hard for many people to collaborate on the code base versus before when the code velocity was slower. And then, as you go outside of engineering, the big one I see is access to data, where you'd like to have your agent go, how many times a day do people at companies around the world say, hey, what's the status of this deal? And that is information that the company knows and should be agentically answerable. And we actually have some cool stuff at Stripe where I was seeing where you can actually answer that pretty well, but with both habits and access to data, and as you get into a bigger company, the permissions engine of who can actually get access to this data, that all needs to be rewritten. And then you get into role definition where, like you were saying, Eng PM design stems a little bit from a prior year, and you may want to, at least in some cases, merge those roles a little bit as AI gets better at all those, since you've got a product doer. Anyway, that's my characterization of in 2026, the models are capable of this, but we're only using them so much. What do you think that adoption of the intelligence looks like? [SPEAKER_02] Look, a lot of us are working on literally what the Gemini teams, the Gemini Enterprise teams, and the anti-gravity teams, they're all precisely working on these problems. This is the roadmap you're talking about, right? And that's literally, we are using it internally, running into these barriers, working past it. So that's the products that are shipping. We are still diffusing it because what you do is people, as part of using it, like if you're the SRE team at Google, you suddenly find portions which you can create an automated workflow. And so that's happening in these parts, right? But doing it more systematically, when you develop skills, how does it get centralized? How is it available to the models and for everyone to use? Identity access controls are real hard problems. And so we are working through those things, but those are the key things which are limiting diffusion to us too, right? And we take security a lot more seriously, and so we have to, right? So that is another layer on top of all these things, the cost of mistakes when you're running these services. And so we have to work through it. But I think because of it, when we solve it, I think we will bring it in a more robust way, which will help. So I feel like we're going through the fixed cost right now, but you will see these jumps of what people are able to do when we bring it outside and others are doing it too. And in a more robust way, the models are improving. Google re-forecasts its business a few times a year, formally, I presume. At least we do at Stripe, where we set a budget for the year and then three times a year we produce a formal re-forecast. And when you think about it, a re-forecast is a moment in time function where you take the state of the business, some of which is in people's heads, but most of which is written down everywhere where it's like, how is this product doing and how is that product doing? Will this deal close? Will that happen? Will this deal close? Will this deal close? Will this deal close? Will this deal close? Will this come the moment in time stage of the business, we put it into a function, and out come the updated numbers for the year. You can imagine an AI doing a fully no-human-in-the-loop forecast. What quarter do you think Google's first fully agentic forecast is? I definitely expect, in some of these areas, 27 to be an important inflection point for certain things, even the people doing it, that is the workflow through which they would produce it. And maybe for a while, you would check it in the conventional way, but you cross over. But I expect 27 to be a big year in which some of those shifts happen pretty profoundly. I think that was Alad's question was Eng is an early adopter, but outside of Eng. [SPEAKER_01] And okay, it sounds like using 27, a lot of these non-Eng processes really start wrapping up. I do think your question earlier on, I think you were asking in the context of way more robotics, companies. I do think companies which are, that's one advantage startups are going to have, more AI native teams. And you can probably get at it through your interview processes, etc. Whereas for us, we would have retraining, transformation, etc. And I think that that's an advantage the younger companies are going to have. And we have to drive the transformation. [SPEAKER_01] Last question. We're talking a lot about initiatives that started small at Google, like the Transformer, which is not Google's main priority when that initiative started. What's a small thing inside Google that you're excited about these days? [SPEAKER_01] It probably would surprise people, when we decided to do data centers in space, we started as a very small team, right? So it's literally a few people with a small budget to go to the first milestone. So I think it's important to start small, even if it's a big idea. So that is an example of a small thing. Look, I literally spent time yesterday with someone who was explaining some improvement in post training, which is like one person talking through the improvement they are doing, listening to it. I'm like, oh, that's going to really show us a nice jump, right? So that's the constant power of this moment. And so all of that, I don't want to be specific about the second one, but we'll publish it one day, I'm sure. But those are some of the small gems I'm excited about. Because it did send us in space and new ML techniques. [SPEAKER_02] Yeah. Yeah. Great answer. Sundar, thank you. All right. Real pleasure. Thanks. Take care. Yeah. Yeah. Yeah. There are a lot of small attributes which can probably make a difference. But I don't think at a foundational level, there is a difference in outlook on what the curve is. Yeah. Right? Or how we internalize the technology. Look, I think even within the company, there's a set of us living on the bleeding edge, firing agents, seeing what these things can do, see the agents pick up skills, do stuff, and also look back three months ago, what they could do now. And we are living that exponential internally. Right? I think you're both right. Where I agree, you can kind of point us at the history of Google. I think what a lot is getting at is like a feeling where I saw a tweet go by that someone was saying, what you have to realize to explain what's currently going on in the valley is that every tech executive has severe AI psychosis right now and is spending, you know, a huge amount of time writing code and talk to AI and things like that. I thought that was a funny take and not without any truth to it. And I'm curious, what were your feeling the AGI moments along the way of the recent, or, you know, to what extent do you have AI psychosis these days? My first feeling the AGI moment was 2012 when Jeff Dean demoed the earliest version of Google Brain. This is when the neural networks recognized a cat, right? So that was 2012. I went with Larry to the DARPA challenge. It might have been 2014, I think I need to be exact about when we went there, seeing the cars drive there. Demoing the earliest versions of the models having what we would call as imagination. So there have been many moments like that. So it was obvious the technology is progressing. In terms of living now and kind of having a visceral feel for it. I think the closest I would say is if you're coding and you give it a complex task and you never open the IDE and you're in some agent manager world and you see it kind of do it, you know, and how powerful it is. So, you know, if you can call it feel AGI. So there are moments like that. Yes, yes. I did a little hobby project recently and after a while I was like, oh, I wonder what language it's using. But that was like a detail that I needed to ask it about after everything was up and running. Yeah, you would just like magic. Yeah. Yeah. So, you know, moments like that for sure. And but the slope of the curve is what surprises you. Yeah. Right. And you're improving it on so many paradigms, it feels clear that there's going to be progress ahead. Right. So... When you talk about the visceral feel, I feel like one thing that's important to tech companies and every CEO thinks about this differently is how you stay connected to the product experience and everyday users because tech products are so abstract that it's easy to, you know, you cannot just manage through reports from teams and slide decks and spreadsheets. And so, you know, Tony Xu is talking about how he still works as a door dasher, you know, to stay very connected to that experience. We do at our like little weekly all hands weaver occurring segment of just walk the store where we like click around in the dashboard together and we're tripping over like why is that modal there and that's a bit confusing or whatever just so we're like collectively using the product. I'm curious how it works for you and how at Google you ensure that you're staying connected to the experience of using the products. Other than you use like Gmail and everything every day. Oh, yeah. You know, like, you know, dogfooting like literally internal versions. I do block time like to kind of use it intensely. So, like kind of focus time to do it. And so, that helps. Like even just two weeks ago, I was stretching in the gym and I had the phone with Gemini Live. And so, I'm like, I want to talk to it for like the entire 30 minutes on like one topic. So, you do those things and some of it works, some of it is frustrating. But you kind of learn a lot, right? Like, so, I force myself to use it in those power user mode ways and stay in touch that way. X helps because sometimes you get the raw feedback. Thank you for fixing the Google Calendar thing. That was so good. Well, there's a few more we have to fix. But thanks for flagging it. So, yeah, X helps because you kind of get the raw comments and I try to follow it directly. But I'll tell you what has helped. Internally, like I would go fire to our earlier part. Like I would query in anti-gravity, just our internal version of anti-gravity. Hey, we launched this thing. Like, what did people think about this? Tell me that like the worst five things people are talking about, the best five things people are talking about. And I type that. Now, that brings it back. So, has my life gotten easier? Yes. So, in the past, I would have to spend a lot more time trying to get a sense for it. Now, an AI agent is helping me in that journey. So, you can get, you know, well, how much should I be spending firsthand to get that feel versus actually leveraging these tools? So, even I'm going through a journey there, right? So, I'm trying to adapt to this future. I guess there's, you mentioned, A, that it's not zero sum. B, there's all these productivity gains people are seeing. And if you look at a lot of prior technology cycles, it took a while for the internet or for mobile or for SaaS to show up in actual GDP numbers, right? In the context of AI, we're seeing it from a data center buildup perspective, right? That's driving part of GDP growth. How do you think ahead in terms of three, four or five years, do you think the US economy is bigger because of AI? And if so, how much bigger? Look, for these returns to make sense, somewhere it has to, you know... How long was it before? I think it was maybe from Sequoia, someone wrote and saying, people are investing this much. Yeah, they're comparing the CapEx to the... Yeah. And this might have been two and a half years ago. And it was a talk and like saying, it doesn't make sense because you would need to return at that level. You're probably 10x things. Yeah. Yeah, yeah, yeah. Since that moment, I need to go look at the numbers again, right? So at some point, it has to reconcile. To be very clear, we are supply constrained. We are seeing the demand across all the surface areas. I actually don't have any doubt that this is a massive market and outcome. So my question... And I think there's a lot of things that people misunderstand. So for example, people often talk about software engineering budgets, and then what proportion of that is token versus salary. And to some extent, I think that market has been so demand constrained for great software engineers that suddenly adding supply can 10x that market, right? In other words, I think the market for software engineering and coding is dramatically bigger than anybody thinks. And it's the wrong metric to say, you know, token budget versus engineers. So I actually think it should grow a lot of things. Yeah. I was just sort of curious of your view of like, how much growth do we think is likely actually to come of this? I actually wasn't doubting at all sort of CapEx versus outcomes or, you know? I see. Yeah. Look, I mean, going back to the internet and looking at GDP growth, you know, it doesn't quite capture what we all feel with the internet, right? And so maybe we would have had negative GDP growth without the internet. Consumer surplus. Yeah. So, you know, it's tough to look ahead. I do think there are natural dampening mechanisms in society at various levels. And the obvious ones being, you know, the compute build out is a different curve than the rate at which we can improve the models, right? So you're already dealing with a more constrained curve there. Then how do you diffuse it into society, right? We are doing this with Waymo, right? And you can make Waymo safer than human drivers, but you have to be careful at like the pace at which we are rolling out, et cetera. So sometimes, you know, how do you diffuse it through society responsibly? There are constraints in all these layers, right? But I think the US economy is so much larger than it was 10 years ago. So to grow that, even at a half a percentage point higher, then, you know, that's a massive contribution. So I expect it to play out that way. Listening to Sundar is a powerful reminder of what it means to operate at true internet scale. When it comes to commerce, though, most businesses are forced to make critical decisions in a vacuum. They view the internet economy through only their own lens, using only the data that exists within their four walls. This is where Stripe comes in. Because we process $1.9 trillion a year in global payments, we have a panoramic view. When you build on Stripe, you're plugging in to that network intelligence. From verifying a user's identity, to stopping fraud before it starts, to streamlining how businesses connect and knowing what payment methods work best locally, we've built a system that puts 1.6% of global GDP to work to protect and grow your revenue. So if you want to use the power of Stripe's network intelligence, come see what you can build here. You referenced the supply constraints. And I think that's a really interesting defining aspect of 2026, basically, where you said 150 billion in CapEx, 180? We have said it'll be between 175 and 185. Okay. So 180-ish billion of CapEx. And what's interesting to me is that Google could not spend $400 billion in CapEx if it wanted to, because the memory isn't there, and the power isn't there, and all these components. So can you just tick through? We can find a number of electricians we would need. Exactly. So I'd love to hear just your overview of the various bottlenecks. Look, at some level, you have to work back to actual wafer capacity or something like that. So there are deeper ground troops. I think so wafer starts, it's kind of a fundamental constraint. I think power and energy are more solvable. Permitting and actually working through a regulatory environment might be a constraint. So the pace at which you can do things. Even though there's lots of land in pro-growth, Texas or Nevada or Montana, just maybe not enough? I think we're making tremendous progress. I think for the US, I think it's a particularly important thing. You're in awe of the pace in China, how fast they can build things. So I really think we need to learn to build things much faster. You almost have to shift your mentality to think about what would it take to do things 10x faster in the physical world? Construct 10x faster. But I would worry about that as a constraint. I think there could be growing resistance. So it's not as simple as a few people deciding you want to build faster. The data center moratorium and stuff. Yeah. Yeah. So I would say wafer starts, the ability to permit and do things. And I do think there's a lot of good work being done from the government on. I think people realize you need to do these things better. Then comes critical competence in the supply chain. Memory is a good one. We are constraining those things in the short term. Everyone will respond to it. But I think all of us running companies, regardless of how AGI-pilled you are, then comes this error bands of like, you know, how bullish can you be? What's the margins you can afford? Because there are extraneous factors which can go wrong in the world, right? Which are outside of control. So everyone is making those adjustments. Those are all constraints, right? And constraints. So I think that's where I see the constraints. Is memory the biggest component that you think about? Memory is definitely one of the most critical components now. Yes. And you said in the short term, do you think just people ramp up supply and so high prices will take care of us? There is no way that the leading memory companies are going to dramatically improve their capacity. So you have those constraints in the short term, but they get more relaxed as you go out. Yes. But I do expect all of this to constraint. By the way, I think it'll push a lot of innovations on, we will make these things 30x more efficient. Yes. Like, so all that is happening simultaneously as well. Does that enforce an oligopoly market? So if you actually look on the model side, because if you look at a lot of the views of models and how they're going to improve, a lot of it is going to be both self-improvement. So the models will start writing more and more pieces of themselves, do more data labeling for themselves, et cetera. So it's a musical chairs game of who has compute right now, basically? Exactly. Who has compute right now and how much can you actually scale relative to overall industry capacity? And if everybody is roughly per rata up to some number, you've effectively put a ceiling on how much far ahead somebody can pull versus everybody else. Do you think that's a correct statement or an incorrect statement? I think it's a reasonable framework to think about it that way. But there are things which are, you know, I'm coming here as we just shipped Gemma for, right? And it's a really good open source model. I mean, the Chinese models are very good. Yeah, but anything outside of China, you know, it's a very good open source model. You know, the frontier to Gemma 4 is both huge and not so huge in terms of time. Like Gemma 4 is based on Gemini 3 architecture, right? You know, it's a very weird thing, right? You're talking about a set of weights, which can fit on a USB stick. Yeah, yeah. It's amazing. Right? So, so it's like a really, you know, crazy. It's not like a SpaceX rocket. Yeah, I guess like, I'm always shocked that you run a data center for months and months and months and then your output is a file. Yeah. Literally, it's like having a Word doc or something and that's your model. It's amazing. So there are these unique attributes about this. So which when, which makes me challenge those frameworks and say, you know, how should we think about this? But I think it's a reasonable, at least on the inference side, what you're saying is a very reasonable way to think about it. Think about it. But I do think, I do think everyone is trying to figure out how to blow through the capitalist incentive to break through these constraints. Yeah. You know, it's immense. But as you say, there's only so much memory in the world. So like, no capitalist incentive will really solve 26 or 27 memory supply. That may be the era where you see more divergence. Yeah. So, you know, and remember that has to balance with wafer capacity increasing, you being able to permit those data centers. So this constraint may be less severe than it appears. Right? So you have to kind of envision the total square set of like all the things that you need and then, and think it through, right? Are you worrying capital? Yes. Yes. But, but again, what's interesting to me is that plausibly people would invest beyond the current capex, but we're now just running against 26 and 27 real world constraints. It's a little about the straight of Hormuz. You can have whatever price of oil you want. Ultimately, if you take 20 million barrels a day out of the system, you need to like destroy 20 million barrels a day of demand. And it's kind of similar with memory where like ultimately some people have to not get the memory they want. But there are other constraints, right? Like which, you know, take security as a constraint. Uh-huh. And these models are definitely like really going to break pretty much all software out there. Maybe already we don't know. Yeah. Yeah. Yeah. I'm not sure if you can hear and speak. Do you really think all software there? Because like SSH, people have been trying to break for a long time. Do you think like just- No, I'm not talking about just think- Just regular software. ... software, large platforms, right? How many zero days? Yeah. Yeah. You know, so there are constraints here in the system, right? You just can't wish away. Right. Somebody was telling me that the black market price of zero days is dropping because the supply is growing due to AI, which I thought was a really interesting- ... market metric. Yeah. Yeah. Not at all surprised, right? And not at all surprised. So, but when- How does it practically diffuse through society? What are the implications of it? Yes. Right? Like, you know, and so there are parallels, I think. So, I think there could be hidden constraints. Yes. And there could be shocks to the system, if you will. But having said that, like, you know, ... I genuinely think there's a lot of upside ahead. Some of the constraints maybe are helpful. Yes. Right? I think constraint inspires creativity. For the compaction cycle where we get more efficient- For the forces maybe important conversations to be had, which otherwise wouldn't happen. Yeah. Right? I think, you know, just on my security point alone, like, I thought about, we are going to need more coordination. Yes. Which is not happening today. There will be a moment of, you know, it could be a sharp moment, right? And like, you know, and so all those things, I don't think you can wish them away. Yes. Yes. Right. Yeah. Actually, related to that, Google does have an amazing portfolio of things that's both built and bought into. From an ownership perspective, you know, you own a reasonable amount of SpaceX, I think. I don't know the exact amount, but I think it was 10-ish percent way back when. Anthropic, 10-ish percent, the majority of Waymo, which is like an amazing thing. And then, internally, obviously, there's this enormous swath of amazing technology that's been developed. We talked about AI and transformers. There's TPUs. Obviously, Waymo was another one of these things. There's Quantum. You know, you just released a very interesting result there. Are there other hidden gems that people should know about or that are especially interesting or that may have very big impact in the future? Or you're peaking people maybe underestimates. Yeah. Yeah. Look, we're constantly trying to take these long-term projects, which, when you first announce them slightly marginally, looks ridiculous. You know, like we're in the earliest stages of thinking about data centers in space, right? But your earlier discussion around constraint inspires creativity. But if you take a 20-year outlook, right, where are you going to put most of these data centers? Really hard problems to solve. But those are examples of projects we think about today, which are Waymo in 2010. Quantum itself is one of these projects. We are, like, in a deeply committed way making progress there. And I'm excited about it. Where do you think Quantum will have the biggest impact? Because mainly people talk about molecular modeling. They talk about cryptography. There's quantum proof sort of cryptography that people have been developing over time. On the molecular modeling side, it actually looks like the deep learning models tend to be very good at that in certain circumstances. I mean, you all pioneered that with AlphaFold. Do you think Quantum will actually matter? And if so, where do you think it'll have the biggest impact? Look at abstract level, to me, it feels like to simulate nature more and more. Like, you know, like given it's inherently quantum, you would need quantum systems to better simulate it. We may get there with classical computing techniques in a surprising way or get at it with enough compression and abstraction. It may work. But I fundamentally felt like Quantum would have an edge there. And I don't know, we still don't understand the Haber process for fertilizer. Like, there are many complex... I mean, you know, it's probably your background going back to what you did in college and what you did in college more. So my, you know, my instinct tells me there'll be, you know, simulating weather, simulating, you know, reality, all that, I think, Quantum 11 advantage. I think the way the history of technology is you get something to a scale where it works and then you use it and people's creativity on the top finds the applications. So, you know, I mean, I always give this example of mobile phones plus GPS enabled Uber. Yeah. Like, like, there's nobody who was working on phones who would predict that as an outcome of this platform shift. So, you know, confident Quantum will have many, many, many applications if you can actually make it work. Yeah. So that's how I think about it. So, sir, we interrupt you. You're talking about kind of your favorite of the Google further afield. I think we're making, you know, the GM team is deeply thinking through robotics, right? And, you know, robotics is an area where we were too early as a company before. It turned out AI was the missing ingredient for a lot of ideas, maybe 15 years ago or 10 years ago. But, you know, the Gemini robotics models are sort of on spatial reasoning, et cetera. So we definitely have state of the art models that's there. And we are partnering back in an ironic way with Boston Dynamics and Agile and a few of the companies and in a determined way making progress. And there are extraordinary startups out there as well. Mm-hmm. But so we are investing in, you know, I spoke about quantum data centers in space, drone delivery with wing. You know, I think we are scaling up wing where in some reasonable time period, like 40 million Americans will have access to a wing delivery service, right? And I'm not talking years out or something like that. But again, these are all like methodical compounding when you take these long term projects. So, you know, we are committed. Isomorphic. Yeah. Isomorphic is very exciting. Yeah. You know, think about, you know, think about being focused on these models in a targeted way to improving all the possible steps in drug discovery. And even though you have long pulls, like phase three trials, et cetera, getting there with a much higher probability of success. Yeah. I think it's definitely the smartest approach I've seen in terms of the different bio models and really thinking about the broader swath beyond just the molecular design, which is, I think, where most of them are stuck. Yeah. It seems very smart. Can I ask, I'm curious, how capital allocation actually works at Google? And what I mean by that is, you know, the idea of good capital allocation is about internalizing that the opportunity cost for capital and putting the cash that a business generates towards its highest invest use. And in the toy example in a business school book, you know, maybe you're Boeing and we can either, you know, we have this cash that our business generates and we can either go bid on the next defense contract and we'll invest this much in R&D dollars and we model this much revenue from the contract, or we go develop a clean sheet commercial airliner and we'll put in this money and we model this kind of thing. It's like a 16% IRR versus a 19% IRR. Okay, I prefer the 19%. In Google's case, the projects are extremely heterogeneous where it's like, okay, we can give the YouTube team more funding so they can go, you know, improve the recommender algorithm and therefore time on site increases and so does monetization. Or we can give the Waymo team more funding so that they can actually get to market faster or scale up faster. Or we can invest in this new AI approach that might, you know, pay off in five years time. And so I'm curious, if you are trying to put capital towards the highest and best use and you're ultimately comparing, how do you compare initiatives that are so different in nature and so different in payoff curve shape? That's the most John question ever. I need to know the answer. You need to throw an ROIC and then it's like- It's a good question. Look, I feel it today more than ever, ironically, because of TPU allocation. Hmm. So in some ways I feel it even way more needs TPUs, right? That's interesting. Yeah. So, you know, computers made the question, ironically, much more front of mind. By the way, of all the things I do, I'm really looking forward to how AI as a companion at least gives inputs to this task. You know, and I think once we can actually get all the data connected and flowing through, I think the models are already capable, it's more, you know, getting all the data unlocked, I think will be helpful. So I feel it there. Historically, I think at Google, one of the advantages we have had is sometimes we make these decisions very early in the cycle. So it's almost like going back to that route, there's a deep technology orientation. And, you know, we actually think about the question you were asking a lot a bit ago about like, what are those longer term things? And so I think thinking at that stage, it's easier because your initial funding amounts can be smaller, but then like, you know, you stay committed for the long term, but you're making sure you're like making progress in a deep way. So as long as you're seeing that underlying technology, like take quantum, for example, how do we judge it? Like we're judging the underlying, like, you know, see if goals around, you know, what logical qubit error-corrected, large stable, logical qubit threshold by when you're going to get to and is the team able to do that? Right? So I think you assess it that way. So one of the, I won't say advantage, I think one of the ways we have thought about it and we've been disciplined about, or at least to me matters a lot is to make those early technology bets in kind of a deep way. And so that's helped. But on a constant basis, look, I've always viewed it as you have to assist the long term value of these things, right? So it's almost like in some intuitive way, you're thinking about the option value and the time of something five to 10 years down the line. And you assume like a crazy growth, right? And think through whether those decisions make sense. So the TPU investments have been great that way, right? And, you know, we've steadily invested in that. Waymo was a great example where I think we increased our investment two to three years ago when the rest of the world got pessimistic on it. When others, some of the people were backing off. It's very magical. It's such a magical experience. I take Waymo now every day to work when I can. And it's magic. So I think Waymo's a good example of this, like this question I have, which is Google does cut projects. And there's various things you've tried where you said, you know, we're actually not going to fund, you know, this part of X all the way, or, you know, we're not going to, you know, we're going to retire this product. It's not working. But Waymo, despite the fact that it was a long road from a compelling demo to commercial service in market, you guys didn't lose the faith. And so what was it that you were seeing? Is that a qualitative decision or a quantitative decision? How do you decide that we're going to cut Loon, but keep Waymo? I think it's to do with that some kind of quantified, you look at the Waymo driver, that's underlying technology, which, you know, how does the software drive the car and the progress in terms of safety and reliability. So it's a long running task, how safe and how will you do it? And you follow that curve and you predict or you set goals where you want to be and how you perform against those curves. I think the team has been phenomenal. There have been maybe phases where it didn't progress, but those are the times you need to kind of like, you know, you have confidence in the quality of the team to break through those phases. But I think the more you're able to evaluate things at that deeper technology level, I think you tend to make those decisions better. Or at least that's how I have tried to do it. One argument I've heard, or one discussion I've heard made about Waymo is that a lot of the huge gains that have been seen recently, because it used to be this hand mapped heuristics of like, how do you deal with edge cases of driving or something happens? How do you respond? And a subset of those were almost like hand drawn out for the cars to follow. And so I kind of a narrow set of things that it could do. And then really, the breakthrough was moving to end to end deep learning a couple years ago as this big transformer wave was happening in general. Do you think if Waymo had been started five years ago, it'd be at the same place as it is relative to having been started 15 plus years ago? Just given that that's the breakthrough that's kind of propelled it forward? Look, I think, you know, we spoke earlier about robotics, you can think about Waymo as a robot, right? I think people who are starting robotics in the last three years, by definition, would be making faster progress, maybe. But I think Waymo is such an integrated system, there are aspects of it, not quite like, you know, like, you know, you take something complex like TSMC or SpaceX launching things, you are talking about system integration in these things in a very complex way. I think Waymo has hidden aspects of that, which the time of how you do it, the craft of it matters. But having said that, I do think the end-to-end approaches are going to be an accident in this series. Because just having a team arguably was a huge benefit to Alphabet and Google, right? I mean, just the fact that you kept investing in it, and then it hit a moment in time where this technology liftoff was more than worth it, and was very smart and forward thinking. I just think it's interesting to ask, how does that apply to other domains? Because to your point on robotics, it seems like robotics will potentially have a different history where you can move very quickly now. Do you folks think about re-internalizing hardware again? Or is it largely going to be a partner-driven model to bringing this stuff to the world? I think we'd keep a very open mind. My lesson from Waymo and on the AI side with TPUs, et cetera, I think to really push the curve well, particularly in areas where you have safety, regulatory, everything. You want the first-hand experience of the product feedback cycle. So I think having first-party hardware will end up being very important, is how I would say right at this stage. Makes sense. So I have two more capital allocation questions. Can you make the case that Google has historically been under-levered, where Google has historically carried a strong-neck cash position? And given that both Google has more ideas than it knows what to do with, like it's just brimming with good ideas, and just the core business grows very durably, and I think Google clearly has a very good understanding of that core business, and it has grown at a higher rate than Google's cost of capital. As you look back on it, should Google have been more leaned in and said, okay, we will be willing to have a leveraged position that's slightly more aggressive than you know, strongly net cash, and we will put that towards new initiatives, or just buy more of this core Google business for Google shareholders, or do more minority investing, which again, Google seems to have been best in class at? It's a great question. For example, if Waymo had reached this point earlier, I think I would have invested the capital earlier. So to some extent, I think you were judging it by, like you want to be good stewards of capital. So to the extent you're bullish on ROIC, you want to invest every last dollar you can there. But to the extent, you know, you have access where you don't think, I mean, this is why we've invested in other companies too, right? Even if not then, but we've always thought about it with the lens of being good stewards of it. We felt our investment in Stripe was being a good steward of our capital. SpaceX, right? You know, SpaceX and Anthropic and so on. So I think now with the AI shift, there are more opportunities on which we can deploy capital in a good way. And so we're doing that. Yes, yes. But I think we always had that mindset. But I would have been glad to invest more capital in Waymo earlier. But we weren't at the level of maturity needed to do that. There was a point in Waymo from a safety standpoint, you know, we did approach Waymo safety first. And you just, it wasn't the right thing to do. So you feel like you cannot point to projects where they would have gone faster had they gone more capital sooner. They just needed a, they had a natural ramp. I wouldn't say that. But I think in generally, at least, we might have gotten the decision wrong, but our approach at least was like to say, if you got excited about something and had the conviction, Yes. We were willing to come in the capital to see through. And my other capital allocation question was, historically at tech companies, the large majority of the R&D expense was the people walking around the building. And, you know, headcount was managed through a very tightly controlled process. And indeed, as you thought about kind of allocating R&D effort, it was really allocating kind of highly paid people to go work on the challenge. And the tech costs were, unless you were doing something very computationally expensive, which obviously Google did in place, you know, Google Books or something. But broadly speaking, the tech was an afterthought compared to the cost of the people. We're now going to a world where, as you say, that's not the case with, you know, TPUs and how you allocate that. Just at a very concrete budgeting level, how does that work inside of Google? Like, are you, do you have an overall TPU budget for the company? And then when you are giving a project resourcing, previously you gave us, you know, a certain headcount budget, and now you give it a headcount and a TPU budget, are they the same budget? Just how does that work when you're doing a quarterly review or an annual review? Look, we've always had a compute budget. Ask me for a friend. Ask me for a friend. Now, we've always had a compute budget, right? You know, even classic compute. I would say with ML, and we use both TPUs and GPUs, by the way, extensively. But ML compute planning is, we are super thoughtful about headcount planning too. But we've always had to plan that. And ML compute, we've gone through phases where they've been easy. And then there have been phases where we've been constrained as a company. But now it is really acutely constrained, right? So you spend a lot more time. I at least spend a dedicated hour a week thinking about that question at a pretty granular level. So I will know by projects and by teams, the compute units they are using, right? And, you know, or at least I have that information, and I'm looking at it and assessing it. And in some ways, it's a really important thing to be doing right now, I feel. How do you learn? So the scarce resource is compute in a lot of cases. And so you're ensuring that Google's precious compute resources are being spent on the most worthwhile initiatives. That's right. Yeah. How do you think about that in the context of GCP and Google Cloud? Because there you're actually allocating the compute to others instead of for your own purposes. And given the constraints in the system, how do you think through that differential allocation? Look, Amy, plan ahead, right? So when we do the forward planning, you know, the cloud team is forward planning and they're putting a plan in place. And, you know, and so you're funding that and you're doing that for our internal needs. You forward plan. And as part of that, you're also signing long-term commitments to customers. Anything we commit to a customer is sacrosanct, right? So these are other contractual commitments. So you solve a lot of it with planning. And so there are, when you plan, we're all in a constrained world. So I think the cloud team would say they don't have the compute they want, et cetera, et cetera. But you solve it with planning ahead. Speaking of Google Cloud, I have my product request that I've been saving up for this section that I know you're looking forward to. You could have posted it on that. Exactly. Yeah, yeah, yeah. But no, I'll say one thing that works really well is the GCP MCP is awesome, where your AI can just interact programmatically with Google Cloud. And I guess you guys have exposed almost everything except like the core, you know, permissioning stuff. And I feel like, in a way, part of the curse of Google Cloud has been there is so much functionality there that I'm sure you occasionally hear from people that it was like a little hard to navigate that you log in, you have to create an organization, a project, and whatever, and find the right services, whatever. And now, all that doesn't matter. And so you just say, you know, hey, go add this Google Cloud functionality. And so that is something that actually, it feels like Google Cloud is really benefiting from like the, it is so broad and there's so much functionality there. I mean, we have a little bit of this problem with Stripe, where as we add more functionality to it, just the right way to navigate this big product surface area is an AI that's read all the API docs for you. So that's working really well. I mean, the promise of AI being this orchestration layer, like for anything you think about, to my earlier question, even internally within the enterprise as a CEO, it's not like you don't have all the data, but how do you get it in one place and you see it in the past that would have meant one more big ERP-ish project to go connect all the data sources, et cetera. Again, like, you know, AI being this orchestration layer in a way that makes sense for the end user, I think it's been delightful to see. So the bigger the product surface area, the more that benefits, you know, hits you. And again, we've seen that to some extent with Stripe, but I feel like with GCP, it must be just a massive effect. I think, I think, I think we could do a lot better. So, but you're right. It's an immense opportunity, I think. Yeah. I've been really happy with it. Okay. And then that gets to my products. Did you bring product suggestions for us? No, you go first. Yeah. I wanted to, but. What's interesting to me about kind of open claw and the product market fit of things like that is they're allowing stateful AI for consumers. And if you want to say, you know, the classic, you know, round up the daily news that I'm interested in and send it to me each morning, or just something that involves persistence, that none of the popular, you know, or like mainstream AI apps allow persistence. Is that common? I think that actually, look, I think you want to give users capability where you have persistent long running tasks. Yes. In a, in a reliable, secure way. Um, you know, you have to think through things like identity, access, et cetera. But I think that's the future. That's the agentic future. Mm-hmm. And bringing that for consumers is like a bit of a exciting frontier we are looking at. Yeah. This is one of mine too. This is, um, Dreamer, which was, uh, the former CTO of Stripes company that just got bought by Meta. I think that a very good version of this. Mm-hmm. It's a very early kind of view of. Yeah. They were making custom software, including persistence, but also, you know, you could kind of spec out. You could kind of make your own little app. Exactly. Yeah. Yeah. And they made that very easy to use. But I feel like when people have this experience, there's a surprise and delight moment. And it's just interesting to me that. Look, I think effectively the consumer interfaces are going to have full coding models underneath, right? And the right harnesses and like the right scales. Yeah. Yeah. And the ability to persist and run somewhere securely in the cloud, locally and in the cloud. So all those primitives are coming together. And so what developers are like today, I feel like there's 1% of the world, maybe not 1%, 0.1% of the world who's kind of living this future. Mm-hmm. Right. Mm-hmm. They are building stuff for themselves. But bringing that to mass adoption Yes. Is a very exciting frontier, I think. Okay. My other product suggestion is, sorry, you have to endure this part of the interview. Awesome. It's the right of passing. Exactly. My other product idea is, for some reason, I don't know if this is your lived experience, but certainly my lived experience, that searching Google Docs is so much harder than, say, search in Gmail. And obviously, they're both equally good search engines. But I think what's going on is keyword search works reasonably well for email, because you can probably remember a unique set of keywords for that email. Whereas what always happens, at least to me, is like, I want to go back and look at the 2026 budget. It turns out if I search Google Slides for 2026 budget, neither of those words is like particularly unique in the context of words that exist in, you know, PowerPoints at Stripe. And so, I can never find the exact right one. And I'm curious, does Sundar Pichai also have this problem? Yeah. Somehow, I haven't felt it as acutely as you're describing it. But when you describe it, it resonates well with my experience. I'm literally playing through the person to whom I'm going to play this segment of the conversation. I know exactly who I'm going to go talk to. The people are working on it. I think we can make it a lot better. I think, look, the AI integration into these services, including Google Docs, I think you will see sharp improvements in the coming months ahead. I think we all did the first versions of it where you just put it in somewhere. But I think, you know, over time, what all can you keep in context? What can you cache? And what can you really bring to bear? I think we can make a lot of progress on. So, I think we can do a lot better. Okay, great. We have a good extra putting up with this. A lot of companies that I'm involved with, even ones that were started reasonably recently, have had to dramatically shift their workflows relative to product development, engineering practices, who they even think of should be on the design team and the capabilities of that. Are you revisiting all that at Google? Are you rethinking it? Has there been big shifts in workflow or other aspects? The way I would say it is, you can think of it as concentric circles. There are some groups within Google who are shifting more profoundly. And so, for me, a big task was, how do you diffuse that to more and more groups, particularly in 2026? Some of it, we couldn't do it early because it breaks so often that, like, you know, it's almost like you see this promising new world, but it's kind of semi-broken. But this year, I feel like the curve is shifting pretty dramatically. So, I can see groups, particularly, I would say, GDM and some of the SWE groups really change their workflows, right? And, you know, they are using, we call this for some strange reason, we have a different name internally than externally of the same product, but it's JetSki internally, which is anti-gravity. And you're living on it, you're living in an agent manager world, you have workflows, and you're kind of working in this new way, right? But just last week, we kind of rolled it out to the search team, right? So, we're constantly pushing that. You know, in a large organization, I think change management is a hard aspect of this technology diffusing, which may be easy for a small company, right? Like, you know, you can quickly switch over. Can I lay out a few problems I see when it comes to actual diffusion of AI in industry? And I'm curious how and when you think we'll solve them? Because I see it with a big intelligence overhang. Like, the AI's are now amazing in terms of what they can do in the abstract. And if you look at how AI native a company is, or just kind of how much it uses that intelligence, there'll probably be a shortfall. And the problems that I see are something like, one, it actually takes a while to get good as an engineer at prompting your AI well. And you can prompt an AI better or worse to write code. Then there's a lot of, say, Stripe-specific prompting in our case to know which tools to use. And so there's kind of the general being good at prompting, and then there's the Stripe being good at prompting. And then, of course, you have the fact that it's hard to share an AI-generated code base because you have a blast radius, and you're just changing so much, and the turnover of the code is high enough, or maybe you're rewriting it several times before you ship, that it's kind of hard for many people to collaborate on the code base versus before when the code velocity was slower. And then, as you go outside of engineering, the big one I see is access to data, where you'd like to have your agent go, how many times a day do people at companies around the world say, hey, what's the status of this deal? And that is information that the company knows and should be agentically answerable. And we actually have some cool stuff at Stripe where I was seeing where you can actually answer that pretty well, but with both habits and access to data, and as you get into a bigger company, the permissions engine of who can actually get access to this data, that all needs to be rewritten. And then you get into role definition where, kind of like you were saying, Eng PM design kind of stems a little bit from a prior year, and you may want to, at least in some cases, merge those roles a little bit as AI gets better at all those, since you've got a product doer. Anyway, that's kind of my characterization of in 2026, the models are capable of this, but we're only using them so much. What do you think that adoption of the intelligence looks like? Look, a lot of us are working on, like, literally what the Gemini teams, the Gemini Enterprise teams, and the anti-gravity teams, they're all precisely working on these problems. This is the roadmap you're talking about, right? Like, you know, and that's literally, we are using it internally, running into these barriers, kind of working past it. So that's the products that are shipping. We are still diffusing it because what you do is people, as part of using it, like if you're the SRE team at Google, you suddenly find portions which you can create an automated workflow. And so that's happening in like these parts, right? But doing it more systematically, when you develop skills, how does it get centralized? How is it available to the models and for everyone to use? Identity access controls are like real hard problems. And so we are working through those things, but those are the key things which are limiting diffusion to us too, right? And we take security a lot more seriously, and so we have to, right? So that is another layer on top of all these things, the cost of mistakes when you're running these services. And so we have to work through it. But I think because of it, when we solve it, I think we will bring it in a more robust way, which will help. So I feel like we're going through the fixed cost right now, but you will see these jumps of what people are able to do when we bring it outside and others are doing it too. And in a more robust way, the models are improving. Google re-forecasts its business a few times a year, formally, I presume. At least we do at Stripe, where we set a budget for the year and then three times a year we produce a formal re-forecast. And when you think about it, a re-forecast is a moment in time function where you take the state of the business, some of which is in people's heads, but most of which is written down everywhere where it's like, how is this product doing and how is that product doing? Will this deal close? Will that happen? Will this deal close? Will this deal close? Will this deal close? Will this deal close? Will this come the moment in time stage of the business, we put it into a function, and out come the updated numbers for the year. You can imagine an AI doing a fully no-human-in-the-loop forecast. What quarter do you think Google's first fully agentic forecast is? I definitely expect, in some of these areas, 27 to be an important inflection point for certain things, even the people doing it, that is the workflow through which they would produce it. And maybe for a while, you would check it in the conventional way, but you kind of switch over, cross over. But I expect 27 to be a big year in which some of those shifts happen pretty profoundly. I think that was Alad's question was Eng is an early adopter, but kind of outside of Eng. And okay, it sounds like using 27, a lot of these non-Eng processes really start wrapping up. I do think your question earlier on, I think you were asking in the context of way more robotics, like companies. I do think companies which are, that's one advantage startups are going to have, more AI native teams. And you can probably get at it through your interview processes, etc. Whereas for us, we would have like retraining, transformation, etc. And I think that that's maybe an advantage like the younger companies are going to have. And we have to, you know, kind of like drive the transformation. Last question. We're talking a lot about initiatives that started small at Google, like the Transformer, which is not Google's main priority, you know, when that initiative started. What's a small thing inside Google that you're excited about these days? Alad Raghuram, It probably would surprise people like, you know, when we decided to do data centers in space, like, you know, we started as a very small team, right? So it's literally a few people with a small budget to go to the first milestone. So I think it's important to start small, even if it's a big idea. So that is an example of a small thing. Look, I literally spent time yesterday who was explaining some improvement in post training, like, which is like one person talking through the improvement they are doing, listening to it. I'm like, oh, that's going to like really show us like a nice jump. Right? So that's the constant power of this moment. And so all of that, I don't want to be specific about the second one, but we'll publish it one day, I'm sure. You know, so, but those are, those are some of the small gems I'm excited about. Because it did send us in space and new ML techniques. Yeah. Yeah. Great answer. Sundar, thank you. All right. Real pleasure. Thanks. Take care.