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

Sundar Pichai is the CEO of Google and Alphabet. He sits down with John and Elad Gil to discuss Google’s resurgence in the AI race, managing a massive $180 billion CapEx budget, and why 2026 is the year of the supply crunch. They cover the constraints of memory and power, why he believes the US economy will grow significantly due to AI, and the internal cultural shift back to “Googley” optimism. Sundar also shares details on long-term bets like data centers in space, why he wishes he had funded Waymo even faster, and the small thing inside Google that still ignites his passion for building.

Timestamps

00:00:18 The history of Google and AI
00:05:17 Speed and Search
00:12:12 Google’s AI comeback
00:27:03 Stripe network intelligence
00:27:53 Bottlenecks
00:41:25 Capital allocation
01:00:44 How Google works

Transcript

[00:00:25.06] John Collison

Sundar 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.

[00:00:26.11] John Collison

Cheers.

[00:00:32.17] John Collison

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?

[00:00:48.17] Sundar Pichai

I think it’s actually worth talking about. It’s a bit misunderstood. Transformers was done in the context of a lot of TPUs, Transformers were all done to solve a specific product need to some extent. The team is thinking about how to make translation better. In the case of TPUs, how do you… Hey, speech rec works, but you suddenly have to sell it to two billion people. We don’t have enough chips for it. How do you solve inference for it?

[00:01:15.14] John Collison

I hadn’t known that. Transformers were specifically—

[00:01:18.19] Sundar Pichai

It was from our research teams, but they were guided by solving product problems. Transformers were immediately used. 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.

[00:01:45.18] Sundar Pichai

We built Transformers and used it immediately in Search to improve language understanding, understanding web pages, understanding your queries, kept building better models. We also started productizing it internally in the form of, there were teams building something called LaMDA. Obviously, we weren’t the first to ship that. But I think it’s less to do with it was just research, and we weren’t applying it in a product direction. That, I think, is just—

[00:02:15.00] John Collison

You did this research, you then saw massive ROI from using it the way you intended, and then you didn’t invent all of the products that were invented with it, but that’s to be expected.

[00:02:24.12] Sundar Pichai

I would go a step further. We exactly even conceived the product, which is ChatGPT. It was LaMDA. If you remember, there was an engineer inside who thought it was sentient. Think of it as an early version of ChatGPT he was speaking to, internally. We even had the product version of it in the multiverse, somewhere else. Google probably shipped that nine months later or something like that.

[00:02:51.11] Sundar Pichai

In fact, in the Google I/O in ‘22, we launched something called AI Test Kitchen, and that was LaMDA, but we had constrained it because internally, we didn’t have an end-to-end version which was RLHF-ed. The version I saw was a lot more toxic at a level. We couldn’t have possibly put it out at that time.

[00:03:16.19] Sundar Pichai

Also, I think 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 were 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. You can look back and say it wasn’t entirely fully obvious.

[00:03:45.01] Sundar Pichai

I think they were lucky to also see it on the coding side with GitHub. I think maybe there was a signal we were missing. Coding side, probably, you were seeing more of a sequential jump than probably just on the language side. Maybe the jumps between GPT-2 and 3 and later 4 were more pronounced if you were using it for coding, too. You can point to things. I think to answer your original question, I think it was less that research-to-product than a bunch of other factors.

[00:04:21.04] Elad Gil

I remember talking to some of the people who worked on ChatGPT, and I think they launched it the week of Thanksgiving. It was a little bit of a buried launch. It wasn’t like, this is a big, prominent thing, and this is going to be an important part of our future.

[00:04:31.12] John Collison

It was clearly a surprise.

[00:04:32.17] Elad Gil

I think it was a cool test case. It was really interesting.

[00:04:34.21] Sundar Pichai

But the way I internalize these moments is if you’re in consumer internet, you’re going to have surprises. We were at Google when, Elad and I, there was something called Google Video Search. YouTube came out. Just that 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.

[00:05:03.08] Sundar Pichai

But the way I’ve internalized is consumer internet, 3 people are going to be sitting and prototyping and throwing out millions of things. I’m not trying to diminish anything, but I’m just saying you’re always going to have these moments. I don’t think people wake up in a garage and ship a better iPhone. That’s not going to happen. But that’s not how consumer internet is. You just have to be conscious of that and to internalize that.

[00:05:31.08] John Collison

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. Original Google Search was really fast and famously displayed the search query time within the results, sort of showing off. Then Gmail Fast Search compared to the competitors of the time, or Chrome, compared to the competitors of the time. And now, I use all of the AI services for different things, but Gemini on TPUs is just so fast. I’m curious how much this is part of the explicit product strategy and how you think of it, or it’s much more nuanced than that.

[00:06:10.12] Sundar Pichai

I’ve always internalized speed. Let’s call it latency for this purpose, and as one of the distinguishing features of a great product. Also, it 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. Both are important.

[00:06:36.18] Sundar Pichai

But you talk about latency. It’s easy to say you want latency, but you’re constantly adding capabilities. The capability frontier is progressing. There’s some sense of, how do you balance that? That’s where it gets more complicated. But to give an example, like Search, I was speaking with the teams. They now have for sub-teams, latency budgets in the milliseconds. You get 50% credit...

[00:07:08.09] Sundar Pichai

If you ship something which shaves off 3 milliseconds, you earn 1. 5 milliseconds for your latency budget, and 1.5 milliseconds gets passed on to the user. 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, but you have rigorous reviews against that. But that’s how much we think it matters.

[00:07:36.19] Elad Gil

For context, I guess humans pick it up in the low hundreds of milliseconds, is that correct? In terms of where it actually impacts?

[00:07:42.12] Sundar Pichai

That’s right. I think we’ve actually, last I checked the dashboard and the metrics, 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 parade of frontier of making sure the capability to speed, and the flash models are at like 90% the capability of the pro models, but much faster, much more effective to serve, and the vertical integration helps and so on.

[00:08:27.05] Elad Gil

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 is incorporated, or Search has incorporated Gemini or AI results in the context of Google. But a lot of people are not 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.

[00:08:46.11] Elad Gil

Instead of asking about trips, it’ll go and plan a trip for you. What do you view as the 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?

[00:08:58.00] Sundar Pichai

I feel like in Search, with every shift, you’re able to do more with it. 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 aren’t looking for web pages, you want to go somewhere, how do you find it? You’re constantly shifting… 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’ll be completing tasks. You’ll have many threads running.

[00:09:35.17] Elad Gil

Will Search exists in 10 years?

[00:09:37.04] Sundar Pichai

Well, you may—

[00:09:38.15] Elad Gil

Or just evolves into something?

[00:09:39.16] Sundar Pichai

It keeps evolving. Search would be an agent manager in which you’re doing a lot of things. I think in some ways, I use Antigravity today, and you have a bunch of agents doing stuff. I can see search doing versions of those things, and you’re getting a bunch of stuff done.

[00:10:00.15] John Collison

I think the root of 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 modality exist?

[00:10:16.20] Sundar Pichai

But today in AI mode in Search, people do deep research queries. That doesn’t quite fit the definition of what you’re saying. But people adapted to that. I think people will do long-running tasks.

[00:10:31.06] John Collison

Yes.

[00:10:31.21] Sundar Pichai

It can be asynchronous.

[00:10:33.04] Elad Gil

We all started, or life started as unicellular organisms, and now we have this complex life. The question is almost like, does that former version or paradigm eventually go away? 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 in the—

[00:10:53.09] Sundar Pichai

I mean, the form factor of devices are going to change. I/O is going to radically change. It’s tough to… I think you can paralyze yourself thinking 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 just do that year ahead. 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.

[00:11:23.07] Sundar Pichai

I think riding the curve itself is exciting. 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. The value of what people are going to be able to do is also on some crazy curve. Once you view it that way, people would ask all these questions. YouTube has done well since TikTok and Instagram has… I can give many examples. I think the more you view it as a zero-sum game, it looks difficult.

[00:12:01.19] Sundar Pichai

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. We are doing both Search and Gemini. They will overlap in certain ways. They will profoundly diverge in certain ways. I think it’s good to have both and embrace it.

[00:12:27.00] John Collison

When we talk about Search and where it’s going and things like this, I’m reminded of the fact that basically a year ago, spring/summer of ‘25, 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, blah, blah, blah. Google was trading for $150-ish a share.

[00:12:49.19] John Collison

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.

[00:13:21.11] Sundar Pichai

It was obviously very invert-focused in that moment. To me, it was very clear in that moment, “Hey, the Overton window shifted.” I felt like the company was built for that moment. The vertical thing, it’s not an accident or something. It was a 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. We were thinking about… The company was operating in an AI-first way. We had deeply internalized this shift.

[00:14:07.08] Sundar Pichai

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. But the exciting part was when I look at it from a 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 like, “Wow, we have this one common technology which can accelerate all those businesses.” Search to YouTube to Cloud to Waymo all relies on progress. It was a very leveraged way to make progress.

[00:14:55.15] Sundar Pichai

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.

[00:15:32.10] John Collison

Was there something that demonstrated to the outside world that, “Oh, they got this”? Was it Gemini 3 that changed people’s minds? I don’t follow all the timelines.

[00:15:39.07] Sundar Pichai

I think the real model, probably, where people saw it was maybe Gemini 2.5. And getting to the frontier on, particularly around multimodality. We made a bunch of… And credit to the Google DeepMind teams. I think we paid a bit more of a fixed cost upfront, but we designed the Gemini models to be very multimodal from day one. There were areas, I think, the strength started showing. Nano Banana was an example of it. You were able to see it all together.

[00:16:20.01] Sundar Pichai

But look, 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, shit. There’s a couple of things left behind.” I think the picture will again be dynamic in a few months. I think the frontier is intense as you would expect it to be. That’s how I think about it.

[00:16:46.10] Elad Gil

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 like the difference between the two or three other labs and the Google team is that Google is not as…

[00:17:02.16] Elad Gil

They call it 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. A, do you think that’s true? And B, do you think that it all impacts some notion of what the future actually looks like and therefore what people are building against?

[00:17:20.12] Sundar Pichai

Look, I think we probably have scaled our CapEx from 30 billion to approximately 180 billion—

[00:17:30.00] Elad Gil

It’s like real money now.

[00:17:32.02] Sundar Pichai

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. I think the founders were AGI-pilled, probably. My earliest conversation… I think this notion that at Google, we haven’t understood what AGI is or Demis and team or Jeff Dean and team. At one point, Demis, Jeff, Ilya, Dario were all there.

[00:18:21.12] John Collison

I like that retort. It’s like, “Hello, have you been paying attention for the past 20 years?”

[00:18:28.00] Sundar Pichai

That doesn’t make sense to me. I think some of it is, if you’re a younger company, or you are more a pure research lab, you may be 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.

[00:18:58.11] Sundar Pichai

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. We are living that exponential internally.

[00:19:19.10] John Collison

I think you’re both right, where I agree, you can point us the history of Google. I think what Elad is getting at is 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 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 feeling the AGI moments along the way of the recent... Or to what extent do you have AI psychosis these days?

[00:19:54.20] Sundar Pichai

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. 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. Demis demoing the earliest versions of the models, having what we would call as imagination.

[00:20:32.11] Sundar Pichai

There have been many moments like that, so it was obvious that 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 ID, and you’re in some agent manager world, and you see it do it, and how powerful it is. You can call it feel the AGI. There are moments like that.

[00:21:03.19] John Collison

I did a little hobby project recently, and after a while, I was like, “Oh, I wonder what language it’s using?” That was a detail that I needed to ask it about after everything was up and running.

[00:21:15.05] Elad Gil

It feels like magic.

[00:21:16.15] Sundar Pichai

Moments like that for sure. The slope of the curve is what surprises you. You’re improving it on so many paradigms. It feels clear that there’s going to be progress ahead.

[00:21:31.07] John Collison

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 cannot just manage through reports from teams and slide decks and spreadsheets. Tony Xu is talking about how he still works as a DoorDasher to stay very connected to that experience.

[00:22:00.00] John Collison

We do our little weekly all-hands, we have a recurring segment of just walk the store where we click around in the dashboard together, and we’re tripping over, why is that modal there? That’s a bit confusing or whatever, just 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 product other than you use Gmail and everything every day.

[00:22:23.22] Sundar Pichai

No, I’m dogfooding, literally internal versions. I do block time to use it intensely, so focus time to do it. That helps. Even just two weeks ago, I was stretching in the gym and I had the phone with Gemini Live. I’m like, I’m going to talk to it for the entire 30 minutes on one topic. You do those things. Some of it works, some of it is frustrating, but you learn a lot. I forced 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.

[00:23:08.14] John Collison

Thank you for fixing the Google Calendar thing. That was so good.

[00:23:11.16] Sundar Pichai

Well, there’s a few more we have to fix.

[00:23:14.17] John Collison

That was awesome.

[00:23:14.23] Sundar Pichai

Thanks for flagging it. Yeah, X helps because you get the raw comments, and I try to follow it directly. I’ll tell you what has helped. Internally, I would go fire... To our earlier part, I would query in Antigravity, just our internal version of Antigravity. “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. Has my life gotten easier? Yes.

[00:23:52.16] Sundar Pichai

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. You can get, well, how much should I be spending first-hand to get that feel? Is this actually leveraging these tools? Even I’m going through a journey there. I’m trying to adapt to this future.

[00:24:11.07] Elad Gil

You mentioned, A, that it’s not zero-sum, B, there’s all these productivity gains people are seeing. 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. In the context of AI, we’re seeing it from a data center build-out perspective. That’s driving part of GDP growth. How do you think ahead in terms of 3, 4, or 5 years? Do you think the US economy is bigger because of AI? If so, how much bigger?

[00:24:37.09] Sundar Pichai

Look, for these returns to make sense, somewhere it has to... How long was it before… I think it was maybe from Sequoia someone wrote, and saying people are investing this much.

[00:24:52.12] John Collison

Yeah, they’re comparing the CapEx to the…

[00:24:55.01] Sundar Pichai

That was two and a half years ago. 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 the investment. Since that moment, I need to go look at the numbers again. 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.

[00:25:20.01] Elad Gil

I actually don’t have any doubt that this is a massive market and outcome. My question, and I think there’re a lot of things that people underestimate. For example, people often talk about software engineering budgets, and then what proportion of that is token versus salary. To some extent, I think that market has been so demand-constrained for great software engineers that suddenly adding supply can 10X that market.

[00:25:40.08] Elad Gil

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. 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.

[00:25:58.05] Sundar Pichai

I see. Look, I mean, going back at the internet and looking at GDP growth, it doesn’t quite capture what we all feel with the internet. Maybe you would have had negative GDP growth without the internet.

[00:26:11.06] John Collison

Consumer surplus.

[00:26:12.01] Sundar Pichai

It’s tough to look ahead. I do think there are natural dampening mechanisms in society at various levels. The obvious ones being the compute build-out is a different curve than the rate at which we can improve the models. You’re already dealing with a more constrained curve there. Then how do you diffuse it into society? We are doing this with Waymo.

[00:26:45.09] Sundar Pichai

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. Sometimes, how do you diffuse it through society responsibly? There are constraints in all these layers. I think... The US economy is so much larger than it was 10 years ago. To grow that even at a half a percentage point higher, then that’s a massive contribution. I expect it to play out that way.

[00:27:23.08] John Collison

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.

[00:27:38.10] John Collison

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.

[00:28:05.16] John Collison

If you want to use the power of Stripe’s network intelligence, come see what you can build here.

[00:28:14.17] John Collison

You reference the supply constraints, and I think that’s a really interesting defining aspect of 2026, basically, where you said 150 billion in CapEx? 180?

[00:28:25.08] Sundar Pichai

We have said it’ll be between 175 and 185.

[00:28:28.20] John Collison

Okay, so 180-ish billion of CapEx. 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. Can you just tick through—

[00:28:44.00] Sundar Pichai

You can’t find a number of electricians we would need.

[00:28:45.22] John Collison

Exactly. I’d love to hear just your overview of the various bottlenecks.

[00:28:50.10] Sundar Pichai

Look, at some level, you have to work back to actual wafer capacity or something like that. There are deeper ground truths. I think so, wafer starts. It’s a fundamental constraint. I think power and energy are more solvable. Permitting and actually working through a regulatory environment might be a constraint. The pace at which you can do things.

[00:29:19.12] John Collison

Even though there’s lots of land in pro-growth, Texas or Nevada or Montana, just maybe not enough.

[00:29:25.07] Sundar Pichai

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. 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. 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 we want to build fast.

[00:30:03.14] John Collison

The data-center moratoriums and stuff.

[00:30:05.06] Sundar Pichai

I would say wafer starts, the ability to permit and do things. 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 components in the supply chain. Memory is a good one. We are constrained in those things in the short term. Everyone will respond to it.

[00:30:30.00] Sundar Pichai

I think all of us running companies, regardless of how AGI-pilled you are, then comes this error band 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. Everyone is making those adjustments. Those are all constraints. That’s where I see the constraints.

[00:31:00.19] John Collison

Is memory the biggest component that you think about?

[00:31:04.18] Sundar Pichai

Memory is definitely one of the most critical components now.

[00:31:08.22] John Collison

You said in the short term, do you think just people will ramp up supply and so high prices will take care of it?

[00:31:14.10] Sundar Pichai

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. I do expect all of this to constrain. I think it’ll push a lot of innovations on. We will make these things 30X more efficient. All that is happening simultaneously as well.

[00:31:51.12] Elad Gil

Does that enforce an oligopoly market? 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 it is going to be both self-improvement. The models will start writing more and more pieces of themselves, do more data labeling for themselves, et cetera.

[00:32:05.09] John Collison

It’s a musical chairs game of who has compute right now, basically.

[00:32:08.11] Elad Gil

Exactly. Who has to compute right now and how much can you actually scale relative to overall industry capacity? If everybody has roughly pro 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?

[00:32:24.18] Sundar Pichai

I think it’s a reasonable framework to think about it that way. There are things which are… I’m coming here as we just shipped Gemma 4. It’s a really good open-source model. The Chinese models are very good, but I think 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. It’s a very weird thing. You’re talking about a set of weights which can fit on a USB stick.

[00:33:03.06] Elad Gil

Yeah, it’s amazing.

[00:33:04.15] Sundar Pichai

It’s like a really crazy... It’s not like a SpaceX rocket.

[00:33:12.23] Elad Gil

I’m always shocked that you run a data center for months, and then your output is a flat file. Literally. It’s like having a Word doc or something, and that’s your model. It’s amazing.

[00:33:21.22] Sundar Pichai

There are these unique attributes about this, which makes me challenge those frameworks and say, “How should we think about this?” 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. I do think everyone is trying to figure out how to blow through the capitalist incentive to break through these constraints. It’s immense.

[00:33:52.11] John Collison

As you say, there’s only so much memory in the world. No capitalist incentive will really solve ‘26 or ‘27 memory supply.

[00:34:02.05] Elad Gil

That may be the era where you see more divergence in models.

[00:34:05.14] Sundar Pichai

Remember, that has to balance with wafer capacity increasing, you being able to permit those data centers. This constraint may be less severe than it appears. You have to envision the total square set of all the things that you need and think it through, including capital.

[00:34:29.07] John Collison

Yes. 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 constraint. It’s a little about this 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. It’s similar with memory where ultimately, some people have to not get the memory they want.

[00:34:57.06] Sundar Pichai

Look, there are other constraints which, take security as a constraint. These models are definitely really going to break pretty much all software out there. Maybe already, we don’t know. We sit here and speak.

[00:35:15.18] John Collison

Do you really think all software there? Because SSH, people have been trying to break for a long time. Do you think…

[00:35:20.02] Sundar Pichai

I’m talking about regular software, large platforms, how many zero-days. There are constraints here in the system. You just can’t wish away.

[00:35:31.19] Elad Gil

Somebody was telling me 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.

[00:35:39.00] Sundar Pichai

Not at all surprised. How does it practically diffuse through society? What are the implications of it? There are parallels. I think there could be hidden constraints, and there could be shocks to the system, if you will. Having said that, I genuinely think there’s a lot of upside ahead. Some of the constraints maybe are helpful. I think constraint inspires creativity.

[00:36:12.12] John Collison

It forces a compaction cycle where you get more efficient.

[00:36:14.23] Sundar Pichai

Forces, maybe. Important conversations to be had which otherwise won’t happen. I think just on my security point alone, I thought about... We are going to need more coordination, which is not happening today. There will be a moment of... It could be a sharp moment. All those things, I don’t think you can wish them away.

[00:36:40.17] Elad Gil

Actually, related to that, Google does have an amazing portfolio of things that it’s both built and bought into. From an ownership perspective, you own a reasonable amount of SpaceX. I don’t know the exact amount, but I think it was 10-ish% way back when, Anthropic, 10-ish%. The majority of Waymo, which is an amazing thing.

[00:36:57.13] Elad Gil

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?

[00:37:17.23] John Collison

Or that you think people maybe underestimate?

[00:37:20.13] Sundar Pichai

Look, we are constantly trying to take these long-term projects, which when you first announced them slightly marginally, it looks ridiculous. We’re in the earliest stages of thinking about data centers in space. Your earlier discussion around constraint inspires creativity. If you take a 20-year outlook, where are you going to put most of these data centers?

[00:37:46.14] Sundar Pichai

Really hard problems to solve. 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.

[00:38:09.04] Elad Gil

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 the AlphaFold. Do you think Quantum will actually matter? If so, where do you think it will have the biggest impact?

[00:38:33.12] Sundar Pichai

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. Abstraction, it may work, but I fundamentally felt like Quantum would have an edge there.

[00:39:05.01] Sundar Pichai

I don’t know, we still don’t understand the Haber process for fertilizer. There are many complex… It’s probably your background going back to what you did in college, more. My instinct tells me there’ll be simulating weather, simulating reality, all that, I think, Quantum will have an advantage. 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 application.

[00:39:41.18] Sundar Pichai

I always give this example of mobile phones plus GPS enabled Uber. There’s nobody who is working on phones who would predict that as an outcome of this platform shift. I’m confident quantum will have many, many applications if you can actually make it work. That’s how I think about it.

[00:40:05.19] John Collison

Sorry, we interrupted you. You were talking about your favorite of the Google further afield initiatives.

[00:40:11.07] Sundar Pichai

I think we’re making a... The GDM 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 or 10 years ago. But the Gemini Robotics models are SOTA on spatial reasoning, et cetera. We definitely have state-of-the-art models there. We are partnering back, in an ironic way, with Boston Dynamics and Agile and a few other companies in a determined way making progress.

[00:40:54.06] Sundar Pichai

There are extraordinary startups out there as well. We are investing in... I spoke about 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 or something like that. Again, these are all methodical compounding when you take these long-term projects. We are committed. Isomorphic.

[00:41:29.04] Elad Gil

Isomorphic is really exciting.

[00:41:31.09] Sundar Pichai

I think about being focused on these models in a targeted way to improving all the possible steps in drug discovery. Even though you have long posts like phase-3 trials, et cetera, getting there with a much higher probability of success.

[00:41:49.01] Elad Gil

I think it’s definitely the smartest approach I’ve seen in terms of the different biomodels and really thinking about the broader swath beyond just the molecular design, which is, I think, where most of them are stuck. It seems very smart.

[00:41:58.21] John Collison

Can I ask, I’m curious, how capital allocation actually works at Google? What I mean by that is the idea good capital allocation is about internalizing the opportunity cost for capital and putting the cash that a business generates towards its highest and best use.

[00:42:20.11] John Collison

In the toy example in a business school book, maybe you’re Boeing, and 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 thing. It’s like a 16% IRR versus a 19% IRR. I prefer the 19%.

[00:42:44.06] John Collison

In Google’s case, the projects are extremely heterogeneous where it’s like, we can give the YouTube team more funding so they can go 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.

[00:43:08.09] John Collison

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?

[00:43:23.19] Elad Gil

This is the most John question ever.

[00:43:25.22] John Collison

I need to know the answer.

[00:43:27.04] Elad Gil

You need to throw an ROIC—

[00:43:30.00] Sundar Pichai

It’s a good question. I feel it today more than ever, ironically, because of TPU allocation. In some ways, I feel it even Waymo needs TPUs. 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.

[00:43:55.08] Sundar Pichai

I think once we can actually get all the data connected and flowing through. I think models are already capable. It’s more of getting all the data unlocked, I think it will be helpful. 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. It’s almost like going back to that roots as a deep technology orientation.

[00:44:21.22] Sundar Pichai

We actually think about the question you were asking Elad a bit ago about what are those longer-term things. Thinking at that stage, it’s easier because your initial funding amounts can be smaller. But then you stay committed for the long term, but you’re making sure you’re making progress in a deep way. As long as you’re seeing that underlying tech... Take Quantum, for example. How do we judge it?

[00:44:47.19] Sundar Pichai

We’re judging the underlying... You have goals around what logical qubit error, correct, a large stable logical qubit threshold by when you’re going to get to, and is the team able to do that?

[00:45:00.00] Sundar Pichai

I think you assess it that way. One of the, I wouldn’t say advantages... 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 a deep way. That’s helped.

[00:45:17.04] Sundar Pichai

On a constant basis, look, I’ve always viewed it as you have to assess the long-term value of these things. It’s almost like in some intuitive way, you’re thinking about the option value and the TAM of something 5–10 years down the line, and you assume a crazy growth and think through whether those decisions make sense.

[00:45:41.01] Sundar Pichai

The TPU investments have been great that way. 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.

[00:45:58.09] Elad Gil

It’s very magical. It’s such a magical experience. I take Waymo now every day to work when I can.

[00:46:02.08] John Collison

I think Waymo is a good example of this question I have, which is Google does cut projects, and there’s various things you’ve tried where you said, “We’re actually not going to fund this part of X all the way, or 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. 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?

[00:46:37.06] Sundar Pichai

I think it’s to do with that, some quantified… You look at the Waymo driver. That’s the underlying technology, which how does the software drive the car? The progress in terms of safety and reliability. It’s a long-running task, how safe and how will you do it. You follow that curve, and you predict, or you set goals where you want to be and how you perform against those curves.

[00:47:03.04] Sundar Pichai

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… You have confidence in the quality of the team to break through those phases. 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.

[00:47:28.20] Elad Gil

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 how do you deal with the edge cases of driving or something happens, how do you respond? A subset of those were almost like hand-drawn out for the cars to follow. It had a narrow set of things that it could do.

[00:47:49.04] Elad Gil

Then really, the breakthrough was moving to end-to-end deep learning a couple of 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.

[00:48:06.12] Sundar Pichai

Look, I think we spoke earlier about robotics. You can think about Waymo as a robot. I think people who are starting robotics in the last three years, by definition, will be making faster progress, maybe. But I think Waymo is such an integrated system. There are aspects of it, not quite like you take something complex like TSMC or SpaceX launching things.

[00:48:31.02] Sundar Pichai

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 case.

[00:48:50.04] Elad Gil

Because just having a team, arguably, was a huge benefit to Alphabet and Google. Just the fact that you kept investing in it, and then it hit a moment in time where this technology lift-off 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 with robotics, we’ll 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?

[00:49:21.06] Sundar Pichai

I think we’d keep a very open mind. My lesson from Waymo and on the AI side with TPUs, et cetera, I need 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. I think having first-party hardware will end up being very important. That’s how I would say right at this stage.

[00:49:52.23] John Collison

Sorry, 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? Given that both Google has more ideas than it knows what to do with, 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.

[00:50:20.02] John Collison

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 strong unit 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?”

[00:50:40.15] Sundar Pichai

It’s a great question. For example, if Waymo had reached this point earlier, I think I would have invested the capital earlier. To some extent, I think you were judging it by... You want to be good stewards of capital. To the extent you’re bullish on ROIC, you want to invest every last dollar you can there. But to the extent you have excess where you don’t think...

[00:51:11.12] Sundar Pichai

This is why we’ve invested in other companies, too, even if not in. But we 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, and Anthropic and so on. I think now with the AI shift, there are more opportunities on which we can deploy capital in a good way, and so we are doing that.

[00:51:41.00] Sundar Pichai

I think we always had that mindset. 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.

[00:52:01.23] John Collison

You feel like you cannot point to projects where they would have gone faster had they gone more capital sooner? They had a natural ramp.

[00:52:10.14] Sundar Pichai

I would say that, but I think in generally, 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, you were willing to commit the capital to see it through.

[00:52:24.17] John Collison

Another capital allocation question was, historically at tech companies, the large majority of the R&D expense was the people walking around the building. Headcount was managed through a very tightly-controlled process. Indeed, as you thought about allocating R&D effort, it was really allocating highly-paid people to go work on the challenge.

[00:52:47.13] John Collison

The tech costs were, unless you were doing something very computationally expensive, which obviously Google did in place of Google Books or something. 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.

[00:53:06.08] John Collison

Just at a very concrete budgeting level, how does that work inside of Google? Do you have an overall TPU budget for the company? Then, when you are giving a project resourcing, previously you gave it 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?

[00:53:30.21] Sundar Pichai

Look, we’ve always had a compute budget.

[00:53:33.09] John Collison

Asking for a friend.

[00:53:34.21] Sundar Pichai

Now, we’ve always had a compute budget, even in classic compute. I would say with ML, and we use both TPUs and GPUs, by the way, extensively. ML compute planning is... We are super thoughtful about headcount planning, too, but we have always had to plan that. 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.

[00:54:03.08] Sundar Pichai

But now it is really acutely constrained. You spend a lot more time. I at least spend a dedicated hour a week thinking about that question at a pretty granular level. I will know by projects and by teams, the compute units they are using, or at least I have that information, and I’m looking at it and assessing it. In some ways, it’s a really important thing to be doing right now, I feel.

[00:54:39.14] John Collison

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—

[00:54:49.10] Sundar Pichai

That’s right.

[00:54:49.23] John Collison

Initiatives.

[00:54:50.20] Elad Gil

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?

[00:55:04.11] Sundar Pichai

Look, we plan ahead. When we do the forward planning, the Cloud team is forward planning, and they are putting a plan in place. You’re funding that, and you’re doing that for our internal needs. You forward plan. As part of that, you’re also signing long-term commitments to customers. Anything we commit to a customer is sacrosanct.

[00:55:30.12] Sundar Pichai

These are contractual commitments. You solve a lot of it with planning. When you plan, we’re all in a constrained world. I think the Cloud team would say they don’t have the computer they want, et cetera. But you solve it with planning ahead.

[00:55:49.09] John Collison

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.

[00:55:55.17] Sundar Pichai

You could have posted it on X.

[00:55:56.14] John Collison

Exactly. 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. I guess you guys have exposed almost everything except the core permission stuff.

[00:56:13.20] John Collison

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. It was 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. Now, all that doesn’t matter.

[00:56:28.22] John Collison

You just say, “Hey, go add this Google Cloud functionality.” That is something that actually it feels like Google Cloud is really benefiting from. It is so broad and there is so much functionality there. 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. That’s working really well.

[00:56:51.12] Sundar Pichai

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 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, AI being this orchestration layer in a way that makes sense for the end user, I think it’s been delightful to see.

[00:57:19.02] John Collison

The bigger the product surface area, the more that benefit hits you. Again, we’ve seen that to some extent with Stripe, but I feel like with GCP, it must be just a massive effect.

[00:57:28.05] Sundar Pichai

I think we could do a lot better. But you’re right, it’s an immense opportunity, I think.

[00:57:33.04] John Collison

I’ve been really happy with it. Then that gets to my product. Did you bring product suggestions for a second?

[00:57:37.00] Elad Gil

You go first. I wanted to but—

[00:57:38.19] John Collison

What’s interesting me about OpenClaw and the product market fit of things like that, is they’re allowing stateful AI for consumers. 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 AI or mainstream AI apps allow persistence. Is that coming?

[00:58:04.22] Sundar Pichai

I think directionally, look, I think you want to give users capability where you have persistent long-running tasks 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. Bringing that for consumers is a bit of an exciting frontier we are looking at.

[00:58:32.07] Elad Gil

This is one of mine, too. This is Dreamer, which was the former CTO of Stripe’s company that just got bought by Meta, I think, did a very good version of this. It was a very early view of...

[00:58:42.23] John Collison

They were making custom software, including persistence, but also you could spec out—

[00:58:49.06] Elad Gil

You can make your own little app.

[00:58:50.09] John Collison

Exactly. They made that very easy to use. I feel like when people have this experience, there’s a surprise and delight moment. It’s just interesting to me that...

[00:59:00.00] Sundar Pichai

Look, I think effectively the consumer interfaces are going to have full coding models underneath, and the right harnesses and the right skills and the ability to persist and run somewhere securely in the cloud, locally and in the cloud.

[00:59:18.00] Sundar Pichai

All those primitives are coming together. What developers are… Today, I feel like there’s 1% of the world, maybe not 1%, 0.1% of the world who’s living this future. They are building stuff for themselves, but bringing that to mass adoption—

[00:59:37.15] John Collison

Yes.

[00:59:38.02] Sundar Pichai

Is a very exciting frontier, I think.

[00:59:41.06] John Collison

My other product suggestion is… Sorry, you have to endure this part of the interview.

[00:59:46.07] Sundar Pichai

It’s the rite of passage.

[00:59:47.08] John Collison

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 search in Google Docs is so much harder than, say, search in Gmail. 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.

[01:00:10.22] John Collison

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. I’m curious, does Sundar Pichai also have this problem?

[01:00:32.14] Sundar Pichai

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 who are working on it. I think we can make it a lot better.

[01:00:51.17] Sundar Pichai

I think 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. 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. I think we can do a lot better.

[01:01:16.13] John Collison

Great. We have a good—

[01:01:18.09] Elad Gil

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 it 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?

[01:01:38.15] Sundar Pichai

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. For me, a big task is 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, it’s almost like you see this promising new world, but it’s semi-broken. But this year, I feel like the curve is shifting pretty dramatically.

[01:02:09.01] Sundar Pichai

I can see groups, and particularly, I would say GDM and some of the SWE groups really change their workflows. 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 Jet Ski internally, which is Antigravity.

[01:02:30.00] Sundar Pichai

You’re living on it, you’re living in an agent manager world, you have workflows, and you’re working in this new way. But just last week, we rolled it out to the Search team. We’re constantly pushing that. In a large organization, I think change management is a hard aspect of this technology diffusing, which may be easy for a small company. You can quickly switch over.

[01:04:56.07] John Collison

Can I add a few problems I see when it comes to actual diffusion of AI in industry? I’m curious how and when you think we’ll solve them. Because as I see it, we have a big intelligence overhang. The AIs are now amazing in terms of what they can do in the abstract. If you look at how AI-native a company is or just how much it uses that intelligence, there’ll probably be a shortfall.

[01:05:21.00] John Collison

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. You can prompt 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. There’s the general being good at prompting, and then there’s the Stripe being good at prompting.

[01:05:42.10] John Collison

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 where 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.

[01:06:01.08] John Collison

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?’” That is like information that the company knows and should be agentically answerable. 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.

[01:06:33.01] John Collison

Then you get into role definition where, like you were saying, Eng, PM, design stems a little bit from a prior year. 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... 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?

[01:07:05.08] Sundar Pichai

Look, a lot of us are working on literally what the Gemini team... The Gemini enterprise teams and the Antigravity teams, they’re all precisely working on these problems. This is the roadmap you’re talking about, right? That’s literally we are using it internally, running into these barriers, working past it. That’s the products that are shipping.

[01:07:34.05] Sundar Pichai

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. That’s happening in these spots. 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. We take security a lot more seriously, and so we have to.

[01:08:13.14] Sundar Pichai

That is another layer on top of all these things, the cost of mistakes when you’re running these services, 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. I feel like we’re going through that fixed cost right now, but you will see this jump of what people are able to do when we bring it outside, and others are doing it, too. In a more robust way, the models are improving.

[01:08:42.03] John Collison

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. 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?

[01:09:03.01] John Collison

How is that product doing? Will this deal close? Will that happen?” Whatever. There’s the moment in time state of the business, we put it into a function and out comes 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?

[01:09:26.23] Sundar Pichai

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. Maybe for a while, you would check it in the conventional way, but you switch over, a crossover. But I expect ‘27 to be a big year in which some of those shifts happen pretty profoundly.

[01:09:54.23] John Collison

I think that was Elad’s question was, Eng is an early adopter, but outside of Eng… Okay, it sounds like you think ‘27, a lot of these non-Eng processes really start—

[01:10:04.05] Sundar Pichai

I do think your question earlier on, I think you were asking in the context of Waymo or 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, et cetera. Whereas for us, we would have retraining, transformation, et cetera. I think that’s maybe an advantage the younger companies are going to have. We have to drive that transformation.

[01:10:38.09] John Collison

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?

[01:10:52.04] Sundar Pichai

It probably would surprise people. When we decided to do data centers in space, we started as a very small team. It’s literally a few people with a small budget to go to the first milestone. I think it’s important to start small, even if it’s a big idea. That is an example of a small thing.

[01:11:13.07] Sundar Pichai

Look, I literally spent time yesterday who was explaining some improvement in post-training, which is one person talking through the improvement they are doing, listening to it, I’m like, “Oh, it’s going to really show up as a nice jump.” That’s the constant power of this moment. 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 jumps I’m excited about.

[01:11:45.06] John Collison

There’s data centers in space and new ML techniques.

[01:11:48.07] Sundar Pichai

Yeah.

[01:11:49.10] John Collison

Great answer. Sundar, thank you.

[01:11:51.07] Sundar Pichai

All right. Real pleasure.

[01:11:52.10] John Collison

Thanks.

[01:11:52.23] Sundar Pichai

Take care.