Satya Nadella — How Microsoft is preparing for AGI

Today, we are interviewing Satya Nadella, we being me and Dylan Patel, who is founder of SemiAnalysis. Satya, welcome.

Thank you, it's great. Thanks for coming over to Atlanta.

Yeah, thank you for giving us a tour of, uh, the new facility. It's been really cool to see. Absolutely. Satya and Scott Guthrie, Microsoft's executive vice president of Cloud and AI, give us a tour of their brand new Fairwater Two data center, the current most powerful in the world.

We try to 10x the training capacity every 18 to 24 months, and so this would be effectively a 10x increase, 10x from what GPT-5 was trained with. And so, to put it in perspective, the number of optics, the network optics in this building, is almost as much as all of Azure across all our data centers two and a half years ago.

It's kind of what 5 million network connections.

You got all this bandwidth between different sites in a region and between the two regions. So is this like a big bet on scaling in the future that you anticipate in the future? There's gonna be some huge model that needs to require 2 whole different regions to train, But

The goal is to be able to kind of aggregate these flops for a large training job and then put these things together across sites. And the reality is you'll use it for, uh, training, and then you'll use it for data gen, you'll use it for inference in all sort of ways. It's not like it's going to be used only for one workload forever.

fairwater four, which you're gonna see under construction nearby, yeah, will also be on that one petabits network. So that we can actually link the two at a very high rate. And then, basically, we do the iWan connecting to Milwaukee, where we have multiple other fairwaters being built.

Literally, you can see the the model parallelism and the data parallelism. It's kind of built for, um, essentially, the training jobs, the pods, the super pods across this campus. And then with the WAN, you can go to the Wisconsin Data Center and literally run a training job with all of them getting aggregated.

And what we're seeing right here is this is a cell with no servers in it yet, no racks.

How many, uh, racks are in a cell?

We think about it, uh, we don't necessarily share that per se, but but we we let me

It's the reason I asked.

well, yeah yeah, uh, you'll see upstairs, I'll start counting, and so you I'll start can start counting. counting. We'll let you start

How many cells are there counting. in this building?

That part also, I can't tell you Yeah, well, division but

is easy, right?

My God, it's Scott, all out.

are you looking at this like now? I see where my money is going.

It's kind of like I run a software company. Welcome to the software company.

How big is the design space once we've decided to use, uh, GB200s and NVLink? How many

Other decisions are there to be made? There is coupling from the model architecture to what is the physical plan that's optimized. And it's also scary in that sense, which is, hey, there's going to be a new chip that'll come out, which obviously, I mean, you take Vera Rubin Ultra. I mean, that's going to have power density that's going to be so different, but with cooling requirements that are going to be so different, right? So, you kind of don't want to just build all to one spec.

So that goes back a little bit to, I think, the dialogue we'll have, which is you want to be scaling in time as opposed to scale once and then be stuck with it.

yeah, When you look at all the past technological transitions, whether it be, you know, railroads or the internet or, you know, replaceable parts and industrialization, uh, the cloud, all of these things, each revolution has gotten much faster in the time it goes from technology discovered to ramp and pervasiveness through the economy. Many folks who have been on Dark Horse's podcast believe this is sort of the final, uh, technological revolution or transition. And this time is very, very different.

Um, and at least so far in the markets, it's sort of, you know, in three years, we've already skyrocketed to, you know, hyperscalers are doing $500 billion of capex next year, which is a scale that's un- unmatched to prior revolutions in terms of speed, and the end state seems to be quite different. How how do you y- y- your your framing of this seems quite different than sort of the, I would say, the AI bro, who is who is quite, um, you know, AGI is coming, and, you know, I I'd like to understand that more.

I mean, look, I I I start with the excitement that I also feel for maybe after the industrial revolution, this is the biggest thing. Um, and so therefore, I I start with that premise, uh, but at the same time, I'm a little grounded in the fact that, um, this is still early innings. Uh, we've built some very useful things, we're seeing some great properties, these scaling laws seem to be working. Um, and I'm optimistic that they'll continue to work, right?

Some of it is, um, you know, s it does require real science breakthroughs, but it's also a lot of engineering and what have you? But that said, I also sort of take the view that, you know, even what has been happening in the last 70 years of computing, uh, has also been a march, uh, that has helped us move, um, you know, with the as I said, y- you know, uh, you know, I. I like one of the things that Raj Reddy, uh, has as a metaphor for what AI is, right?

He's a he's a Turing Award winner out of, uh, CMU. Um, and he's always, and he, he had this even pre-AGI, uh, but he had this metaphor of, uh, AI should either be a guardian angel or a cognitive amplifier. I love that. Uh, it's a simple way to think about what this is ultimately, what is its human utility. It is going to be a cognitive amplifier, uh, and a guardian angel. And so if I sort of view it that way, I view it as a tool.

But then you can also go very mystical about it and say, well, ah, this is, you know, more than a tool. It does all these things which only humans did so far, but that has been the case with many technologies in the past. Only humans did a lot of things, and then we had tools that did them.

I guess, uh, we don't have to get wrapped up in the definition here. But maybe one way to think about it is like, maybe it takes 5 years, 10 years, 20 years. At some point, eventually, a machine is producing satya tokens, right? And the Microsoft board thinks that satya tokens are worth a lot.

How much, how much are you wasting of this, uh, of of, like, economic value by interviewing Satya?

Yeah, you could not afford the API costs of satya tokens. Um, but so, uh, you know, whatever you wanna call it, is that are the satya tokens a tool or an agent? Whatever? Um, right now, if you have models that cost on the order of dollars or cents per million tokens, there's just an enormous room for expansion. Uh, a margin expansion there, where satya a million tokens of satya are, like, worth a lot. Um, and where does that margin go, and what level of that margin is Microsoft involved in?

Is the question I have.

So I think, um, in s in some sense, this goes back again to essentially what's the, uh, economic growth picture gonna really look like, um, what's the firm gonna look like, what's productivity gonna look like? And that, to me, is where, again, if the industrial revolution created after whatever 70 years of diffusion is when you started seeing the economic growth, right? It took, that's the other thing to remember is, um, even if the tech is diffusing fast, uh, this time around, for true economic growth to appear, it has to sort of diffuse to a point where the work, the work artifact and the workflow has to change.

And so that's kind of one place where I think, uh, the change management required for a corporation to truly change, I think, is something we shouldn't discount. So I think going forward, do humans and the tokens they produce, uh, get higher leverage, right? Uh, whether it's the Dwarkesh or the Dylan tokens of the future. I mean, think about the amount of technol- would you be able to run SemiAnalysis on this podcast without technology? No chance, right?

I mean, the scale that you have been able to achieve, no chance. So the question is, what's that scale? Is it gonna be 10xed with something that comes through? Uh, absolutely, uh. And therefore, whether you're ramped to some revenue number or you're ramped to some audience number or what have you. And so that, I think, is what's going to happen, right? I mean, the the point is, uh, that's whatever. What took 70 years, maybe 150 years for the industrial revolution, may happen in 20 years, 25 years.

That's a better way to, like, I would love to compress what happened in 200 years of the industrial revolution into 20 year period, if you're lucky.

So Microsoft historically has been perhaps, you know, the greatest software company, the largest software as a service company. You know, you've gone through a transition in the past where you used to sell Windows licenses and disks of Windows or Microsoft, and now you sell, you know, subscriptions to 365 or, um, a a, as as we go from sort of, you know, that transition to where where your business is today, um, there's also a transition going after that, right?

Uh, software as a service, incredibly low incremental cost per user. Uh, there's a lot of R&D, there's a lot of customer acquisition cost. This is why not Microsoft, but the SaaS companies have underperformed massively in the markets because the COGS of AI is just so high, and that just completely breaks how these business models work. H- How do you, as a as a, as perhaps the greatest software company, um, software as a service company, transition Microsoft to this new age where COGS matters a lot, um, and and the incremental cost per user is different, right?

Cause right now you're charging, hey, it's twenty bucks for Copilot.

Yeah, so I think that this is a it's a great question because, in some sense, the business models themselves, I think the levers are gonna remain similar, right? Which is if I look at the the if you if you look at the menu of models, uh, starting from, like, say, consumer all the way right. There will be some ad unit, uh, there will be some transaction, there'll be some device gross margin for somebody who builds an AI device.

Um, uh, there will be subscriptions, consumer and enterprise, uh, and then there'll be consumption, right? So I still think that that's kinda how those are all the meters. To your point, what is a subscription? Up to now, people like subscriptions because they can budget for them, right? They are essentially entitlements to some consumption rights that come encapsulated in a subscription. So that, I think, is what, in some sense, it becomes a pricing decision. Uh, so how much consumption is in you are entitled to is if you look at all the coding subscriptions, that's kinda what they are, right?

And then you kind of have the pro tier, the standard tier and what have you? And so, I think that's how the pricing will, uh, you know, and the margin structures will get tiered. Um, the interesting thing is having at Microsoft, the good news for us is we kinda are in that business, uh, all, you know, across all those meters. And in fact, that at a as a portfolio level, uh, we pretty much have consumption subscriptions, uh, to all of the other consumer levers as well.

Um, and then I think time will tell which of these models make sense in what categories. Um, one thing on the SaaS side since you brought up, which I think a lot about, is, uh, take, uh, Office 365 or Microsoft 365. I mean, man, uh, having a low ARPU is great because here here's an interesting thing, right? During the transition from server to cloud, one of the questions we used to ask ourselves is, Oh my God, if all we did was just basically move the same users who were using, let's call it, our Office licenses and our servers at that time, office servers, right, uh, to the cloud, and we had cogs.

This is going to basically not only shrink our margins, uh, but we'll be fundamentally a non-profitable or a less profitable company. Except what happened was the move to the cloud expanded the market like crazy. Uh, right, I mean, we sold a few servers in India, didn't sell much, whereas in the cloud, suddenly everybody in India also could afford fractionally buying, uh, servers, the IT cost. I mean, in fact, the biggest thing I had not realized, for example, was the amount of money people were spending buying storage underneath SharePoint.

In fact, EMC's biggest segment may have been storage servers for SharePoint. All that sort of dropped in the cloud because nobody had to go buy. In fact, it was working capital. I mean, basically, it is cash flow out, right? And so it expanded the market massively. So this AI thing will be that, right? So if you take coding, um, what we built with GitHub and VS Code in over whatever decades, uh, suddenly, the coding is that big in one year?

And so that, I think, is what's gonna happen as well, which is the market expands massively.

I I guess there's a question of the market will expand, will the parts of the revenue that touch Microsoft expand? So, copilot is an example where, if you look, uh, early this year, I think, uh, I guess, according to Dylan's numbers, um, the copilot revenue, GitHub copilot revenue was like $500 million or something like that. And then, uh, there were, like, no close competitors, whereas now you have Claude Code, Cursor, and Copilot with around similar revenue around a billion, and then Codex is catching up around $700, $800 million.

And so the question is, across all the services that Microsoft has access to, what is the advantage that Microsoft's equivalents of Copilot have?

Yeah, by the way, I love this chart. You know, I love this chart for so many reasons. One is we're still on the top. Um, second is all these companies that are listed here are all companies that have been born in the last 4 5 years. Yeah, that to me is the best sign, right? Which is if you have new competitors, new existential problems, when you say, Man, who's it now? Oh, Claude's gonna kill you, Cursor is gonna kill you.

It's not Borland, right? So thank God, like, that means we are in the right direction, but this is it, right? The fact that we went from nothing to this scale Is the market expansion. So this is like the cloud like stuff, this, uh, fundamentally, this category of coding and AI is probably going to be one of the biggest categories, right? It is a software factory category. In fact, it may be bigger than knowledge work. So I kind of want to keep myself open-minded about, I mean, we're gonna have tough competition.

I think that's your point, which I think is a great one. Uh, but man, like, I'm glad we have we parlayed, uh, what we had into this, and now we have to compete. And so in the compete side, uh, even in the last quarter, I we just fin- we did our quarterly, uh, announcement. I think we grew from 20 to 26 million subs, right? So I feel good about our sub growth, is, uh, and where the direction of travel on that is.

But the more interesting thing that has happened is, guess where all the repos of all these other guys, uh, who are generating lots and lots of code go to? They go to GitHub. So it GitHub is at an all-time high in terms of repo creation, PRs, everything. So that, in some sense, we wanna keep that open, by the way, that means we wanna have that, right? Because we don't wanna conflate that with our own growth, right?

The f- interestingly enough, we are getting 1 developer joining GitHub a second or something that is the stat, I think. And then 80 of them just fall into some GitHub Copilot, uh, workflow just because there are, and by the way, many of these things will even use some of our coding, uh, code review agents, which are by default on just because you can use it. So we'll have many, many structural shots at this. The thing that we are also gonna do is what we did with Git, get the primitives of GitHub, whether starting with Git to issues to actions.

These are powerful, lovely things because they kind of are all built around your repo. So we wanna extend that. At yeah, last week at GitHub universe, that's kind of what we did, right? So we said, agent HQ was the conceptual thing that we said we are gonna build out. This is where, for example, you have a thing called mission control, and you go to mission control. And now I can fire off. Sometimes I describe it as the cable TV of all these AI agents, because I'll have essentially packaged into one subscription Codex, Claude, um, you know, cognition stuff, anyone's agents, Grok, all of them will be there.

So I get one package, and then I can literally go issue a task, steer them, so they will all be working in their independent branches. Uh, I can monitor them, uh, so I literally have, because I think that's going to be one of the biggest places of innovation, right? Because right now, I wanna be able to use multiple agents, I wanna be able to then digest the output of the multiple agents. I wanna be able to then keep a ha- a handle on my repo.

So it's there's coming some kind of a heads-up display that needs to be built. And then for me to quickly steer and triage what the coding agents have generated. That to me, between VS Code, GitHub, and all of these new primitives we'll build, uh, f- as mission control, I think, uh, with a control plane observability, I mean, think about every, uh, one who's gonna deploy all this will require a whole host of observability of what agent did what at what time to what code base.

So I feel that's the opportunity. Uh, and at the end of the day, your point is well taken, which is we better be competitive and innovate, and if we don't, yes, we will get toppled. But I like the chart, at least as long as we're on the top, even with competition.

The key point here is sort of that GitHub will keep growing irregardless of whose coding agent wins. But that that market only grows at, you know, call it 10, 15, 20%, which is way above GDP. It's a great compounder, but these AI coding agents have grown from, you know, call it $500 million run rate at the end of last year, which was basically just GitHub Copilot, to now the current run rate across, you know, GitHub Copilot, Claude Code, Cursor, Cognition, Windsurf, Replit, uh, Codex, OpenAI Codex.

That's that's that's run rating at five, $6 billion now, um, for the for the Q4 of of this year. That's a 10x, right? And and when you look at, hey, what's the TAM of of software agents? Is it is it the $2 1 trillion of wages you pay people? Or is it is it is it something beyond that? Uh, because every company in the world will now be able to, you know, develop software more.

Yeah, so I think that this is a it's a great question because, in some sense, the business models themselves, I think the levers are gonna remain similar, right? Which is if I look at the the if you if you look at the menu of models, uh, starting from, like, say, consumer all the way right. There will be some ad unit, uh, there will be some transaction, there'll be some device gross margin for somebody who builds an AI device.

Um, uh, there will be subscriptions, consumer and enterprise, uh, and then there'll be consumption, right? So I still think that that's kinda how those are all the meters.

No question Microsoft takes a slice of that, but you've gone from near 100% or certainly way above 50% to, you know, sub 25% market share in just one year. What is the sort of confidence that people can get that Microsoft I mean, there's will be?

no again, uh, it goes back a little bit to sort of there's no birthright here that we should have any confidence other than to say, hey, we should go innovate. And knowing Yep, the the lucky break we have in some sense is that, uh, yep, this category is gonna be a lot bigger than anything we had high share in. Let's let me say it that way, right? In some sense, you could say, man, we kind of had high share in VS Code, we had high share in the repos for with GitHub, uh, and that was a good market.

But the point is, even having a decent share in what is a much more expansive market, right? I mean, you could say we had a high share in client-server server computing, we have much lower share than that in hyperscale, but is it a much bigger business by orders of magnitude? So, at least there's existence proof that Microsoft being okay, uh, even if our share position yep, has not been as strong yep, as it was, uh, as long as the markets we are competing in are creating more value, uh, and there are multiple winners.

Uh, so I think that's the stuff. But I w- I, I take your point that ultimately it all means you have to get competitive. So I watch that every quarter, and so that's why I think, wha- I'm very optimistic that, uh, what we are going to do with GitHub HQ and or agent HQ, turning GitHub into a place where all these agents come. Uh, as I said, we'll have multiple shots on goal on there, right? It need it need not be that, hey, some of these guys can succeed along with us.

Uh, and so it ne- doesn't need to be just one winner, uh, and one subscription.

I I I guess the reason to focus on this question is that it's not just about GitHub, but fundamentally about Office and all the other software that Microsoft offers. Which is that one vision you could have about how AI proceeds is that, look, the models are k- going to keep being hobbled, and you will need this direct, visible, um, observability all the time. And another vision is over time, these models can now they're doing tasks that take 2 minutes.

In the future, they'll be doing tasks, next, they're gonna be doing tasks that take 10, 30 minutes. In the future, maybe they're doing days worth of work autonomously. And then the model companies are charging thousands of dollars, maybe for access to, uh, really, a coworker, which could use any UI to communicate with their human and, uh, so forth, and migrate between platforms. So, w- if we're getting closer to that, why aren't the model companies that are, uh, just g- getting more and more profitable, the ones that are taking all the margin?

Why is the the place where the scaffolding happens, which becomes less and less relevant as AI's become more cap capable, gonna be that important? And that goes to, you know, Office as it exists now versus coworkers that are just doing knowledge work

I think that's a great point. autonomously. So I mean, I think that's a g- I mean, for example, I mean, this is where, you know, does all the m value migrate just to, uh, the model, um, and, uh, or does the mo- you know, the uh, does it get split between the scaffolding, um, and uh, the model and what have you? I I think that, uh, time will tell, but my my fundamental point also is the incentive structure gets clear, right?

Which is if you take, um, let's take, uh, let's take information work, but take even coding. Uh, already, in fact, one of the favorite settings I have, uh, in GitHub Copilot, is called Auto, um, right? Which will just optimize, in fact, I buy a subscription, the auto one will start picking and optimizing for what I am asking it to do, uh, and it could even be fully autonomous, and it could sort of arbitrage the tokens available across multiple models to go get a task done.

So if that is the m that that means that if you take that argument, the commodity there will be models, uh, and especially with open-source models, you can pick a checkpoint and you can take a bunch of your data and you're seeing it, right? I think all of us will start, you'll whether it's from Cursor or from Microsoft, uh, we'll start seeing some in-house models, even, uh, which will, and then you'll offload most of your, uh, tasks to it.

So I think that one argument is if you win the scaffolding, uh, which today is dealing with all the hobbling problems or the, uh, the jaggedness of these intelligence problems, which you kinda have to. Um, if you win that, then you will vertically integrate yourself into the model, just because you will have the liquidity of the data and what have you, and there are enough and more checkpoints that are going to be available. Uh, that's the other thing, right?

So structurally, I think there will always be an open source model, uh, that will be fairly capable in the world that you could then use as long as you have something that you can use that, uh, with, which is data, uh, and a scaffolding, right? So I can make the argument that, oh my God, uh, if you're a model company, you may be, uh, you may have a winner's curse. You may have done all the hard work, done n- unbelievable innovation, except it's kinda like one copy, uh, away from that being commoditized.

And then the person who has the data for grounding and context engineering, um, and the liquidity of data can then go take that checkpoint and train it. So I think the argument can be made both ways.

Unpacking? Sort of what you said, there's two views of the world, right? One is that models, there's so many different models out there, open source exists. There will be differences between the models that will drive some level of, you know, who wins and who doesn't, but the scaffolding is what enables you to win. The other view is that actually models are the key IP, and yes, we're in a very everyone's in a tight race, and there's some, you know, hey, I can use Anthropic or OpenAI, and you can see this in the revenue charts, right?

Like, OpenAI's revenue started skyrocketing once they finally had a code model similar capabilities to Anthropic, although d in different ways. Um, th- th- there's the view that, like, the model companies are actually the ones that garner all the margin, right? Because, you know, if you look across this year, at least on Anthropic, their gross margins on inference went from, you know, well below 40% to north of 60%, right? By the end of the year. Um, the these the margins are expanding there despite, hey, more Chinese open source models than ever.

Hey, OpenAI's competitive, hey, Google's competitive, hey, X Grok is now competitive, right? All these, all these companies are now competitive, and yet, despite this, the margins have expanded at the model layer significantly. Um, h- how do you think about the

it's a g- it's a great question. I mean, it, it, I I think the the one thing is perhaps a few years ago, people would say, Oh, I can just wrap a model and build a successful company. Uh, and that, I think, is g- probably gotten debunked just because the model capabilities, um, and with tools use in particular. But the interesting thing is there's no like when I look at Office 365, let's take even this little thing we built called Excel Agent.

It's interesting, right? Excel agent is not a UI level wrapper, it's actually a model that is in the middle tier, uh, in this case, because we have all the IP from the the GPT family. Uh, we are taking that and putting it into the core middle tier of the Office system to both teach it what it means to natively understand Excel, everything in it. So it's not just, hey, I just have a pixel level understanding. I have a not full understanding of all the native artifacts of Excel, uh, both when I see it.

Like, because if you think about it, if I'm gonna give it some reasoning task, right? I need to even fix the reasoning mistakes I make. And so that means I need to both not just see the pixels, I need to be able to see, oh, I got that formula wrong, and I need to understand that. And then, so, to some degree, that's all being done, not at the UI wrapper level with some prompt, but it's being done in the middle tier by teaching it all the tools of Excel, right?

So I'm giving it even essentially a markdown to teach it the skills of what it means to be a sophisticated Excel user. So it's a weird thing that it it goes back a little bit to AI brain, right? Which is you're building not just Excel, you are now b- business logic in its traditional sense. You're taking the Excel business logic in the traditional sense and wrapping essentially a cognitive layer to it using this model, which knows how to use the tool.

So, in some sense, Excel will come with an analyst bundled in and with all the tools used. That's the type of stuff that'll get built by everybody. So even for the model companies, they'll have to compete, right? So if their price stuff high, uh, guess what? If I'm a builder of a tool like this, I'll substitute you. I may use you for a while. And so as long as there's competition, there's always a winner-take-all thing, right?

If there's won't be one model that is better than everybody else with massive distance, yes, that's a winner-take-all. As long as there's gonna be competition where there's multiple models, just like hyperscale competition, and there's an open-source check, uh, there is enough room here, uh, to go build value on top of models, uh. But at Microsoft, the way I look at it and say is, uh, we are going to be in the hyperscale business, which will support multiple models.

We will have access to open AI models for, uh, uh, you know, 7 more years, which we will innovate on top of. So royalty f- I mean, essentially, I think of ourselves as having a frontier class model, uh, that we can use and innovate on with full, uh, flexibility, and we'll build our own models, uh, with MAI, um, and and so we will always have a model level. And then we'll build these, whether it's in security, whether it's in knowledge work, whether it's in coding or in science, we will build our own application scaffolding, which will be model forward, right?

It won't be a wrapper on a model, but the model will be wrapped into, uh, the application.

I have so many questions about the other things you mentioned, but before we move on to those topics, um, I still wonder whether this is, like, not forward-looking on AI capabilities, where you're imagining models like they exist today, where, yeah, I can you have this like it takes a screenshot of your screen, but it can't, like, look inside each cell and what the formula is. And I think the better m mental model here is like, look, a human, just imagine that these models actually will be able to actually use a computer as well as a human.

And a human knowledge worker who is using Excel can look into the formulas, can y- you know, use alternative software, can migrate data between Office 365 and another piece of software if the migration is necessary, et cetera. So what is

That's kind of what I'm saying. but but but So what, what?

if that's the case, then the integration with Excel doesn't matter that No, no, I much, eh? Because

Yeah, no, I I just I don't worry then about the Excel integration. yep, after A all, Excel was built as a tool for analysts. Great, So whoever great. is this AI that is an analyst should have tools that And they can they can use.

in the computer, right? The the just the way a human can use a computer, that's their tool.

th- th- th- The tool is the computer, right, all right. So tha- so all I'm saying is I'm building an analyst as as essentially an AI agent, uh, which happens to come with, um, an a priori knowledge of how to use all of these analytical tools.

but is it is it something, uh, maybe just, uh, just to make sure we're talking about the same thing. Um, is it a thing that a hu- like me using Excel as a podcaster, I'm not proficient

No, no, it would be in, it, but, an autom- no, no, like, a completely autonomous. So just imagine I work, it's like, so it's we should now maybe yeah, sort of lay out how I think yeah, the future of the company is right. Uh, the future of the company would be the tools business, which I have a computer, yeah, I use Excel, and in fact, in the future, I'll even have a co-pilot, right, uh, and that co-pilot will also have agents, right?

That's still I yeah, am I, you know, it's still me steering everything yeah, and everything is coming back. So that's kind of one world. Then the second world is the company just literally provisions a computing resource for an AI agent, and that is working fully autonomously, that fully autonomous agent will have essentially embodied set of those same tools, uh, that are available to it, right? So this AI tool that comes in also has not just a raw computer, uh, because it's going to be more token efficient to use tools to get stuff done.

In fact, I kinda look at it and say, our business, which today is an end-user tools business, will become essentially an infrastructure business in support of agents doing work. Is another way to think about it, right? So, uh, if one of the things that you'll see us do, and in in fact, like all the stuff we built underneath M365, still is going to be very relevant. Uh, you need some place to store it, some place to do archival, some place to do discovery, some place to manage all of these activities, even if you're an AI agent.

So that's so it's kind of a new infrastructure.

s so just just to make sure I understand, you're saying, like, look, theoretically, a future AI that has actual computer use, which is all these companies are working on, the model companies are working on right now, could use, even if it's not partnered with Microsoft or under our umbrella, could use Microsoft software. But you're saying we're gonna give them w- if if you're working with our infrastructure, we're gonna give you, like, lower level access. That makes it more efficient for you to do the same things you could otherwise do anyways.

100%. I mean, so the entire thing, I in in fact, the way, the, you know, like, what happened is we had servers, then there was virtualization, and then we had many more servers. So that's another way to think about this, which is, hey, don't think of the the tool as the end thing. What is the entire substrate underneath that tool that humans use? And that entire substrate is the bootstrap for the AI agent as well, because the AI agent needs a computer.

That's kind of one, like, you know? So, in fact, one of the fascinating things we are seeing a significant amount of growth is all these guys who are doing these, office uh, artifacts and and what have you as autonomous agents and so on, want to provision Windows 365, right? They really want to be able to provision a computer for these agents. Uh, and so absolutely, and that's where I think we are gonna have essentially an end-user computing yeah, infrastructure business, which I think is going to just keep growing because, guess what, it's going to grow faster than the number of users.

So, in fact, that's kind of one of the other questions people ask me is, Hey, what happens to the per-user business? yeah, At least the early signs yeah, may be the way to think about the per-user business is not just per-user, it's per-agent. And if you sort of say it's per-user and per-agent, the key is what's the stuff yeah, to provision for every agent, a computer, um, a set of security things around it, an identity around it, uh, and all those things, observability and so on are the management layers, and that's, I think, all going to get baked into that.

the way to frame it, at least the way I currently think about it, and I'd like to hear your, you know, your view, is that, uh, these model companies are all building environments to train their models to use Excel or Amazon shopping or whatever, whatever it is, book flights. Um, but at the same time, they're also training these models to do migration from, because that that is probably the most immediate, uh, valuable thing, right? Converting mainframe-based systems to standard cloud systems, converting, um, Excel databases into real databases, uh, with SQL, right?

Or or converting, um, you know, wh- what is done in Word and Excel to something that is more programmatic and more efficient in a classical sense that can actually be done by humans as well. It's just not cost effective for the software developer to do that. That seems to be what everyone is going to do with AI for the next, you know, few years, at least, to massively drive value. Um, h- how does Microsoft fit into that if the models can utilize the tools themselves to migrate to something?

And yes, Microsoft has, you know, a leadership position in databases and in storage and and and all these other categories. But the use of, say, a office ecosystem is going to be significantly less, just like potentially the use of a mainframe ecosystem could be potentially less. Now, mainframes have grown for the last two decades, actually, even though no one talks about them anymore, they've still grown. yeah, 100%. I Uh, h- how does

agree with that.

how does that flow forward

Yeah, I mean, uh, at

now?

the end of the day, this is not about sort of, hey, um, there is going to be a significant amount of time where there's gonna be a hybrid world, right? Because people are gonna be using the tools, they're gonna be working with agents that have to use tools, and by the way, they have to communicate with each other. What's the artifact I generate that then a human needs to see? So, like, all of these things will be real considerations in any place.

So the outputs, inputs. So I don't think it'll just be about, oh, I migrated off, right? But the bottom line is I have to live in this hybrid world. So let, but that doesn't fully answer your question because there can be a real new efficient frontier where I just agents working with agents, uh, and completely optimizing. And even when agents are working with agents, what are the primitives that are needed? Uh, do you need a storage system?

Uh, does that storage system need to have e-discovery? Does that e-discovery? Uh, do you need to have observability? Do you need to have an identity system that is going to use multiple models with all having one identity system? So these are all the core underlying rails we have today for what are office systems or what have you? Uh, and that's what I think we will have in the future as well. You talked about databases, right? I mean, take, uh, you know, man, I would love all of Excel to have a database backend, right?

In fact, I would love for all that to happen immediately. Uh, and that database is a good database. I mean, databases, in fact, will be a big thing that'll grow. Uh, in fact, if I think about all of the office artifacts, uh, being structured better, the ability to do the joins between structured and unstructured better because of the agentic. What? That'll grow the underlying, what is infrastructure business? It happens, the consumption of that is all being driven by agents.

You could say all that is just in time generated software by a model company. that could also be true if we we will be one such model company too. Uh, and so we will build in. So the competition could be, uh, that we will build a model plus all the infrastructure and provision it, and then there will be competition between a bunch of those folks who can do that.

Um, I guess, speaking of model companies, you say, okay, we will also be one of the not only we'll have the infrastructure, we'll have the model itself. Right now, Microsoft AI's most recent model that was released two months ago is 36 in Chatbot Arena. And there's a qu- I mean, uh, you obviously have the IP rights to OpenAI, so there's a question of first, to the extent you agree with that, it seems to be behind. Why is that the case, especially given the fact that you could, um, you theoretically have the right to just, like, fork OpenAI's mono repo or distill on their models?

Um, y- yeah, es- especially if it's a big part of your strategy that we need to have a leading model company.

Yeah, I mean, f- so, first of all, we are absolutely going to use the OpenAI models, uh, to the maximum, uh, across all of our products, right? I mean, that's, I think, the core thing that we're gonna continue to do all the way for the next 7 years, uh, and not just use it, uh, but then add value to it. That's kind of where the analyst and this Excel agent, and these are all things that we will do, where, you know, we'll do, uh, you know, uh, RL fine-tuning.

We'll do some mid-training runs on top of a GPT family where we have unique data assets and build capability. Um, the MAI model, the way I think we're gonna think about it is the the good news here, in fact, with the new agreement, is even we can be very, very clear that we are gonna build a world-class superintelligence team and go after it with high ambition. But that at the same time, we're also gonna use this time to be smart about how to use both these things.

So that means we will, on one end, be very product focused, on on the other end, be very research focused. In other words, uh, because we have access, uh-huh, to the GPT family. The last thing I don't wanna do is use my flops in a way that is just duplicative and doesn't add much value. So I wanna be able to take, uh, the flops that we use to generate a GPT family and maximize its value, while my MAI flops are being used for, let's take the image model that we launched, which I think this launched, uh, is at number 9 in the, uh, image arena.

You know, we're using it, you know, both for cost optimization. It's on Copilot, it's in Bing, and we're gonna use that. We have a a audio model in Copilot, which it's got personality and what have you? We optimized it for our product, so we will do those. Even on the LLM arena, we started on the text one, I think it was it debuted at night 13. And by the way, it was v- it was done only on whatever 15,000, uh, H100s.

And so it was a very small model. And, uh, so it was again to prove out, uh, the core capability, the instruction following and everything else, which, but, you know, we wanted to make sure we can match what was state of the art. And so that shows us, given scaling laws, what we are capable of doing if we gave more flops to it, right? So the next thing we will do is an omni model, where we will take sort of the work we have done in audio, what we have done in image and what we have done in text.

That'll be the next pit stop on the MAI side. So when I think about the MAI roadmap, we're gonna build a first-class superintelligence team. We're gonna continue to drop and do on in the open some of these models. They will either be in our products being used because they're going to be latency friendly, COGS friendly, or what have you, or they'll have some special capability. And we will do real research in order to be ready for some next 5, 6, 7, 8 break breakthroughs, uh, that are all needed on this march towards superintelligence.

So I think that's and while exploiting the advantage we have of having the GPT family that we can work on top of as well.

Say, we roll forward 7 years, uh, you no longer have access to OpenAI models. What does one get confidence, or what does Microsoft do to make sure they are leading, uh, have a leading AI lab, right? Today, you know, it's it's all OpenAI has developed many of the breakthroughs, whether it be scaling or reasoning, or Google's developed all the breakthroughs like transformers, uh, but but it it. It is also a big talent game, right? You know, you've seen Meta spend, you know, north of $20 billion on talent, right?

Uh, you've seen Anthropic pr- uh, poach the entire BlueShift reasoning team from Google last year, you've seen Meta poach a large reasoning and post-training team from Google more recently. These these sorts of talent wars are very capital intensive. They're the ones that, you know, arguably, you know, if you're spending $100 billion on infrastructure, you should also spend, you know, x amount of money on on the people using the infrastructure so that they're more efficiently making these new breakthroughs.

What what confidence can one get that, you know, hey, Microsoft will have a team that's world-class that can make these breakthroughs? And, you know, once you decide to turn on the money faucet, you know, you're you're being a bit capital efficient right now, which is which is smart, it seems, uh, to not waste money, doing duplicative work. But once you decide you need to, you know, how how can one say, Oh, yeah, now you can shoot up to where the top 5 models

Well, look, I mean, at the e- at the end of the day, we're gonna build a world-class team, and we are already have a world-class team that's beginning to be sort of assembled, right? With Mustafa coming in, we have Karen, we have Amar, Subramanian, who did a lot of the post-training at Gemini 2.5, who's at Microsoft, Nando, who did a lot of the multimedia work at DeepMind is there. And so we're gonna build a world-class team.

And in fact, I think later this week, even Mustafa will publish some, you know, a little more clarity on what our lab is going to go do. Um, I think the thing that I want, uh, the world to know, perhaps, uh, is we are gonna build the infrastructure that'll support multiple models, uh, you know, uh, we. Because from a hyperscale perspective, we wanna build the most scaled infrastructure fleet that's capable of supporting all the models the world needs, whether it's from open source or whether, obviously, from open AI and others.

And so that's kind of one job. Second is in our own model capability. We will absolutely use the open AI model in our products, and we will start building our own models, and we may, like in in GitHub Copilot, Anthropic is used. So we will even have other frontier models that are going to be wrapped into our products as well. So I think that that's kind of how, at least each time, at the end of the day, the eval of the product as it meets a particular task or a job is what matters.

And we'll sort of back from there into the vertical integration needed, uh, knowing that as long as your service, you know, you're serving the market well with the product, you can always cost optimize.

There there's a question going forward. So right now, we have models that have this distinction between training and inference, and one could argue that there's like a s smaller and smaller difference between the different models. Um, going forward, if you're really expecting something like human level intelligence, humans learn on the job. You, you know, if you think about your last 30 years, what what makes satya tokens so valuable? It's the last 30 years of wisdom and experience you've gained at Microsoft.

Um, and we will eventually have models, if they get to human level, which will have this ability to continuously learn on the job. And that will drive so much value to the model company that is ahead, at least in my view. Because you have copies of one model broadly deployed through the economy, learning how to do every single job. And unlike humans, they can amalgamate their learnings to that model. So there's this sort of continuous learning, sort of exponential feedback loop, um, which almost looks like a sort of intelligence explosion.

Uh, if that happens, and Microsoft isn't the leading model company by that time, doesn't then this, uh, you know, you're saying, well, we substitute one model for another, et cetera. Matter less, cause it's just like this one model knows how to do every single job in the economy, the other long tail don't.

Yeah, no, I think that your point about if there's one model that is the only model that's most broadly deployed in the world and it sees all the data and it has continuous learning, that's game, set, match, and, you know, it's shut shop, right? I mean, the the reality, at least I see, um, is the world, even today, for all the dominance of any one model, it's not the case. Um, it, it, like, take any, take coding.

There's multiple models. In fact, every day, it's less the case where there is not one model that is getting deployed broadly. In fact, there's multiple models that are getting deployed. It's kind of like databases, right? It's always the thing is like, hey, can one database be the one that just is used everywhere? Except it's not. Uh, there are multiple types of databases that are getting deployed, uh, for different use cases. So I think that there is going to be some network effects of continual learning or data.

You, you know, I'll call liquidity that any one model has. Uh, is it gonna happen in all domains? I don't think so. Is it gonna happen in all geos? I don't think so. Is it gonna happen in all segments? I don't think so. It'll happen in all categories at the same time, I don't think so. So therefore, I feel like the design space is so large, uh, that there's plenty of opportunity. But your fundamental point is having a capability, which is at the infrastructure layer, model layer, and at the scaffolding layer.

And then to be able to compose these things, not just as a vertical stack, but to be able to compose each thing for what its purpose is, right? You can't build an infrastructure that's optimized for one model. If you did that, what if you go fall behind? In fact, all the infrastructure you built will be a waste, right? You kind of need to build an infrastructure that's capable of supporting multiple sort of families and lineages of models.

Otherwise, the capital you put in, which is optimized for one model architecture, of that means you're one tweak away from some MOE-like breakthrough that happens for somebody else, and your entire network topology goes out of the window. Then that's a scary thing, right? So therefore, you kind of want the infrastructure to support whatever may come, in fact, in your own model family and other model families, and you got to be open. If you're, if you're serious about the hyperscale business, you've got to be serious about that, right?

Um, if you are serious about being a model company, you've got to basically say, Hey, what are the ways people can actually do things on top of the model? So that I can have an ISV ecosystem unless I'm thinking I'll own every category, that just can't be that. Then you won't have an API business, and that, by definition, will mean you'll never be, uh, a platform company that's gonna be successfully deployed everywhere, right? So, therefore, the industry structure is s such that it will, uh, really force people to specialize.

And that in that specialization, a company like Microsoft should compete in each layer by its merits, uh, but not think that this is all about all a road to game set match, where I just compose vertically all these layers, that's that just doesn't happen.

So, according to Dylan's numbers, there's gonna be half 1 trillion in AI capex next year alone, and labs are already spending billions of dollars to snag top researcher talent. But none of that matters if there's not enough high-quality data to train on. Without the right data, even the most advanced infrastructure and world-class talent won't translate into end value for the user. That's where Labelbox comes in. Labelbox produces high-quality data at massive scale, powering any capability that you want your model to have.

It doesn't matter whether you need a coding agent that needs detailed feedback on multi-hour trajectories, or a robotics model that needs thousands of samples on everyday tasks, or a voice agent that can also perform real-world actions for the user, like booking them a flight. To be clear, this isn't just off-the-shelf data. Labelbox can design and launch a custom production-scale data pipeline in 48 hours, and they can get you tens of thousands of targeted examples in weeks.

Reach out at labelbox.com dwarkesh. All right, back to Satya.

So last year, Microsoft was on on path to be the largest infrastructure provider, uh, by far. You were early 2023, so you were you went out there, you acquired all the resources in terms of leasing data center, starting construction, securing power, everything. You guys were on pace to beat Amazon in 2026 or 2027. Um, but certainly by 2028, you were gonna beat them. Um, since then, you, you know, in, let's call it the second half of last year, Microsoft did this big pause, right?

Where they let go of a bunch of leasing sites that they were gonna take, which then Google, Meta, um, Amazon, in some cases, Oracle, uh, took these sites. We're sitting in one of the largest data centers in the world, so obviously it's not everything. You guys are expanding like crazy, uh, but there are sites that you just stopped working on. Why? Why did you do this, right?

Yeah, I mean, the fundamental thing we this goes back a little bit to what is the hyperscale business all about, right? Which is one of the key decisions we made was that if you're gonna build out Azure to be fantastic for all sort of stages of AI, uh, from training to mid-training to data gen to inference, we just need fungibility, uh, of the fleet. Um, and and so that entire thing caused us not to basically go build, uh, a whole lot of capacity with a particular set of generations, uh.

Because the other thing that you gotta realize is having actually for now up to now 10x every 18 months enough training capacity for the various open AI models. Uh, we realized that, um, the key is to stay on that path. But the more important thing is to actually have a balance to not just train, but to be able to serve these models all around the world. Because at the end of the day, the rate of monetization is what then will allow us to even keep, uh, funding.

And then the infrastructure was going to need us to support, as I said, multiple models and what have you. So once we said that that's the case, since then, we just course-corrected to where the path we are on, right? If I look at the path we are on is we are doing a lot more starts now. Uh, we are also buying up as much capacity as we can, whether it's to build, whether it's to lease, or even GPUs as a service.

But we are building it for where we see the demand, uh, and the serving needs and our training needs. And we didn't want to just be a hoster for one company, uh, and have just a massive book of business with one customer. That that's not a business, right? That is sort of, uh, you know, you should be vertically integrated with that company, uh. And so, given the, uh, the thing that OpenAI was going to be a successful independent company, which is fantastic, right?

I think it's makes sense, right? And even Meta may use third-party capacity, but ultimately, they're all going to be first-party, uh, for anyone who has large scale, they'll be, you know, they'll be a hyperscaler on their own. And so, to me, was to build out a hyperscale fleet and our own research compute, uh, and that's what the adjustment was. Um, and, uh, and then, and so I feel very, very good. Oh, by the way, the other thing is I didn't want to get stuck with massive scale of one generation.

I mean, we just saw the the GB200s, I mean, the GB300s are coming, right? And by the time I get to Vera Rubin, Vera Rubin Ultra, guess what? The data center is gonna look very different because the power per rack, power per row, is going to be so different. Uh, the cooling requirements are going to be so different, and that that means I don't want to just go build out like a whole number of gigawatts that are only for a one generation, one family.

And so I think the pacing matters, and the fungibility, and the location matters, uh, the workload diversity matters, customer diversity matters, and that's what we're building towards. The other thing that we have learned a lot is, um, every AI workload does require not only the AI accelerator, but it requires a whole lot of other things, right? And in fact, a lot of the margin structure for us will be in those other things. And so therefore, we want to build out Azure as being fantastic for the long tail of the workloads, because that's the hyperscale business, while knowing that we've got to be super competitive, starting with the bare metal for the highest-end training.

And but that can't crowd out the rest of the business, right? Because we're not in the business of just doing five contracts with five customers, being their bare metal service. That's not a a Microsoft business. That may be a business for someone else, and that's a good thing. What we have said is we are in the hyperscale business, which is, at the end of the day, a long-tail business, uh, for AI workloads. And in order to do that, we will have some leading bare metal as a service capabilities for a set of models, including our own, uh, and that, I think, is the balance you see.

The another sort of question that comes around this whole fungibility topic is, okay, it's not where you want it, right? You would rather have it in a good population center like Atlanta, as, y- we're here. Um, there there's there's also the question of, like, well, how much does that matter if, as the horizon of AI tasks grows? Well, actually, you know, 30 seconds for a reasoning prompt, or, you know, 30 minutes for a deep research, or, you know, it's gonna be hours for software agents at some point, um, and days and so on, and so forth.

The time to human interaction. Why does it matter if it's Yeah, it's if it's,

a great, it's a great uh, a location question.

A, B or C?

That's exactly right. So, in fact, that's one of the other reasons why we wanna think about, like, hey, what is an Azure region look like, and what is the, in fact, the networking between Azure regions? So this is where, uh, I think, as the model capabilities evolve, and I think the usage of these tokens, whether it's synchronously or asynchronously evolves. And, in fact, you don't want to be out of position, right? Then, on top of that, by the way, what are the data residency laws, right?

Where do I like? I mean, the entire EU thing, uh, for us, where we literally had to create an EU data boundary, uh, basically meant that you can't just round-trip a call to wherever, even if it's asynchronous. And so, therefore, you need to have maybe regional things that are high density and then the power costs and so on. But you are 100% right in bringing up that the topology, as we build out, uh, a will have to evolve.

One for tokens per dollar per watt. Uh, what are the economics? So, uh, overlay that with what is the usage pattern, um, usage pattern in terms of synchronous, asynchronous, but also, what is the compute storage? Because the latencies may matter for certain things. Uh, the storage better be there. If I have a Cosmos DB close to this for session data, or even for an autonomous thing, then that also has to be somewhere close to it, and so on.

So, I think that all of those considerations is what will shape, uh, the hyperscale business.

You know, prior to the pause, you were you were you, you know, versus, you know, what we had forecasted for you by twenty-eight, you're gonna be, like, twelve, thirteen gigawatts, and now we're at, you know, nine and a half or so, right? But, you know, something that's even more relevant, right? And it's it's, you know, I just want you to, like, more concretely state that this is the business you don't wanna be in, but, like, Oracle's going from, like, one-fifth your size to bigger than you by end of 2027.

And while it's not a Microsoft-level quality of return on invested capital, right? They're still making 35% gross margins, right? Sort of the question is, like, does it is it isn't it is it is it you know, hey, it's not Microsoft's business to maybe do this, wha- w- but you've created a hyperscaler now by refusing this business, by giving away the right of, Uh, first refusal, I I'm Et cetera.

Not, eh first of all, I don't I don't wanna take away any, uh, uh, thing from the success Oracle Has had, uh, in building that business, and I wish them well. And so, the thing that I think I've answered for you is, it didn't make sense for us, uh, to go be a hoster for one model company, uh, with limited time horizon RPO. Let's let's just put it that way, right? The thing that you Have to think through is not what you do in the next five years, but what do you do for the next fifty, uh, because that's kind of what I w- we made our set of decisions, um, I feel very good about our OpenAI, uh, partnership

and what we're doing. We have a decent book, a book of business. We wish them a lot of success. In fact, we are buyers even of Oracle capacity. We wish them success. But, you know, at This point, I think the industrial logic for what we are trying to do is pretty clear, which is, it's not about, like, chasing. Uh, first of all, I track, by the way, your, uh, things, whe- whether it's the AWS or the Google and ours, which I think is super useful.

Uh, but doesn't mean I gotta chase those. Uh, I have to chase them for Not just, uh, the gross margin that they may represent in a period of time. You know, does might what what is this book of business that Microsoft uniquely can go clear, which makes sense for us to clear, and that's what we'll do.

I I I guess I have a question, even stepping back from this of, okay, I I take your point that it's a better business to be in, all else equal, to have a long tail of customers you can have higher margin from, rather than just serving bare metal to a few labs. But then there's a question of, okay, which way is the industry evolving? And so if we believe we're on the path to smarter and smarter AIs, then why isn't the shape of the industry that the open AIs and Anthropics and DeepMinds are the platform which the long tail of enterprises are actually doing business with, where they need bare metal, but, like, they are the platform.

What is the long tail that is directly using Azure, um, because, you know, you you wanna use the general But those models are Purpose?

All gonna be available on Azure, right? So any workload that says, Hey, I wanna use, um, you know, some open source model and an open AI model, like, I mean, if you go to Azure Foundry today, you have all these models that you can provision, buy PTUs, get a Cosmos DB, get a SQL DB, get some storage, get some compute. That's what a real workload looks like. A real workload is not just, Hey, I called an API call to a model.

A real workload needs all of these things to go build an app, uh, or instantiate an application. In fact, the model companies need that, right, to build anything. It's just not like I have a token factory. I have to have all of these things. That's the hyperscale business, uh, and it's not a a any one model, but all these models. And so, if you want Grok plus, let's say, uh, OpenAI plus an open source model, come to Azure Foundry, provision them, build your application.

Here is a database. That's kind of what the business is. Uh, you c- there is a separate business called just selling raw bare metal services to model companies, and that's the argument about how much of that business you wanna be in and not be in, and what is that? It's a very different segment of the business, which, uh, we are in, and we also have limits to how much of it is going to crowd out the rest of it.

Uh, but that's kind of, at least, the way I look at it.

So so, there's there's sort of two questions here, right, like, why why couldn't you just do both is one, and then the other one is, um, given, you know, our our estimates on what your capacity is in 2028 is three and a half gigawatts lower, sure, you could have dedicated that to open AI training and inference capacity, but you could have also dedicated that to, Hey, the this three and a half gigawatts is actually just running Azure, is running Microsoft three six five, it's running GitHub Copilot.

It doesn't actually I could have built it and not given it Or or To OpenAI.

I mean, wanna build it in a different location. I mean, wanna build it in UAE, I may wanna build it in India, I may wanna build it in Europe, right? So, one of the other things is, as I said, like, where we have real capacity constraints right now are given the regulatory needs and the data sovereignty needs, we gotta build all over the world. Uh, it's first of all, state-side capacity is super important, and we're gonna build everything.

But one of the things is, when I look out to twenty thirty, uh, I have sort of a global view of what does Microsoft's shape of business by first party and third party, third party segmented by the frontier labs and their how much they want versus the inference capacity we wanna build for multiple models, um, and our own research compute needs, right? So, that's all what's going into my calculus versus saying, hey, um, I think you're rightfully pointing out the pause, but the pause was not done because we said, Oh my god, we don't wanna build that.

We realized that, oh, we wanna build what we wanna build slightly differently, uh, by both workload type as well as geo-type and timing as well. Like, we'll keep ramping up our gigawatts, and the question is, at what pace and in what location and in what sort of how do I write even the Moore's law on it, right? Which is, do I really wanna overbuild three and a half in twenty seven, or do I wanna spread that in twenty seven, twenty eight, knowing even one of the biggest learnings we had even with NVIDIA is their pace increased, uh, in terms of their model mi- I mean, their migrations.

So, that was a big factor. I didn't want to go get stuck for four years, five years of depreciation on one, uh, generation, and I wanted to just basically buy, like, in fact, Jensen's advice to me was two things. One is, Hey, get on the speed of light execution. That's why I think even the execution in this Atlanta data center I mean, like, ninety days, right, from between when we get it and to hand off to a real workload.

That's sort of real speed of light execution on their front, and so, I wanted to get good on that, and then that way, then I'm building this each generation scaling, uh, and then every five years, then you have a much more balanced so it becomes really literally like a flow, uh, for a large-scale industrial operation like this, where you suddenly are not lopsided, where you built up a lot in one time, and then you take a ma- massive hiatus because you're stuck with all this to your point in one location, which may be great for training, may not be great for inference because I can't serve even if it's, like, it's all asynchronous, but Europe ain't gonna let me round-trip

to u- Texas. So, that's all of the things.

How do I rationalize this statement with what you've done over the last few weeks? You've announced deals with Iris Energy, um, with Nebius, um, and Lambda Labs, and there's a few more coming as well. Uh, you're you're going out there and securing capacity that you're renting from the Neoclouds, um, rather than having built it yourself. What was the what was

I think it's f- it's fine for us because we now have you know, when you have line of sight to demand which can be served where people are building it, it's great. In fact, we'll even have I would say, you know, we will take leases, we will take build-to-suit, we'll take even GPUs as a service where we don't have capacity, but we need capacity and someone else has that. Uh, and by the way, I would even sort of welcome every Neocloud to just be part of our marketplace.

Uh, because again, g- guess what? If they if they go bring their capacity into our marketplace, that customer who comes through Azure will use the Neocloud, which is a great win for them, and we'll use compute, storage, databases, all the rest from Azure. So, I'm not at all thinking of this as just a, you know, Hey, I should just go gobble up all of that myself.

Hm. Um, uh, so you mentioned the co- how the you know, you're depreciating this asset that's five, six years, and this is the majority of the, you know, 75% of the TCO of a data center, and Jensen is taking 75% margin on that. So, what all the hyperscalers are trying to do is develop their own accelerator so that they can reduce this overwhelming cost for, um, uh, equipment to increase the margins. Yeah, and then and then, like, you know, when you look at where they are, right, Google's way ahead of everyone else, right, they've been doing it for the longest.

They're gonna make something like five to 7 million chips, right, of their own TPUs. You look at Amazon, they're trying to make three to 5 million. Uh, but when we look at what, you know, Microsoft is is ordering of their own chips, it's it's it's way below that number. Um, you've had a program for just as long. What's going on with your internal

Yeah, it's a good question.

Chips?

So so the couple of things. One is the thing that is the biggest competitor for any new accelerator is kind of even the previous generation of NVIDIA, right? I mean, in a fleet, what I'm gonna look at is the overall TCO. So, the bar I have even for our own and which, by the way, uh, you know, I was just looking at the data for Maya 200, which looks great, um, e- e- except that one of the things that we learned even on the compute side, right, which is, we had a lot of Intel, then we introduced AMD, and then we introduced Cobalt, and so that's kind of how we scaled it, and so we have good, um, sort

of existence proof of, at least in core compute, on how to build your own silicon and then manage a fleet where all three are at play in some balance, uh, because, by the way, even Google's buying NVIDIA, and so is, uh, Amazon. And makes sense because NVIDIA is innovating, and it's the general purpose thing, all models run on it, uh, and customer demand is there. Because if you build your own vertical thing, you better have your own model, uh, which is, you know, either gonna use it for training or inference, and you have to generate your own demand for it or subsidize the demand for it.

So, therefore, you wanna, uh, make sure, um, you scale it appropriately. So, the way we are gonna go do it is, um, have a closed loop between our own MAI models and our silicon, because I feel like that's the re that that's what gives you the birthright to really do your own silicon, right, where you literally have, uh, designed the microarchitecture with what you're doing, and then you keep pace with your own models, and, um, in our case, the the good news here is OpenAI has a program, uh, which we have access to, um, and so, therefore, to think that Microsoft is not going to have something that's What level of Scalable

Access do you have to that?

All of it.

You just get the IP for all of that?

That absolutely.

So, the only IP you don't have is a consumer hardware.

That's it. Yeah. Yeah. No, and by the way, uh, we gave them, uh, a bunch of IP as well to bootstrap them, right? So, this is one of the reasons why they had a massive because we built all these supercomputers together, uh, or rebuilt it for them, and they, uh, benefited from it, rightfully so, and, uh, and now, as they innovate even at the system level, we get access to all of it. Uh, and, uh, we first wants to in in wanna instantiate what they build, uh, for them, uh, but then we'll extend it, and so to think that we don't have and so, if anything, the way I I think about to your question is, uh, Microsoft wants to

be a fantastic I'll call it speed of light execution partner for NVIDIA because, quite frankly, that fleet, uh, is life itself. I'm not worried about I mean, obviously, Jensen's doing super well with his margins, but the TCO has many dimensions to it, and I wanna be great at that TCO. Uh, d on top of that, I wanna be able to sort of really work with the OpenAI lineage, uh, and the MAI lineage and the system design, knowing that we have the IP rights on both ends.

Hm, uh, spe- speaking of rights, one thing, you know, you you had an interview a couple days ago, uh, where you said that we have rights to the the new agreement you made with OpenAI, you have rights the exclusivity to the stateless API calls that OpenAI makes, and we were sort of confused about if there's any state whatsoever, I mean, you were just mentioning a second ago that all these complicated workloads that are coming up are gonna require memory and databases and storage and so forth, and is that now not stateful?

If ChatGPT is storing stuff on So, but that's the Sessions

Reason why. So, the the thing, the business, the strategic decision we made and also accommodating for the flexibility OpenAI needed in order to be able to procure compute for essentially, think of OpenAI having, um, a PaaS business and a SaaS business. SaaS business is ChatGPT, their PaaS business is their API. That API is Azure exclusive. The SaaS business, they can run it anywhere.

And they can partner with anyone they want to to build SaaS products.

The so, if they want a partner and this partner wants to use a s a stateless API, then Azure is the place where they can get the stateless API.

It seems like there's a way for them to make you you you know, build the product together and and it's a stateful No, even that, Thing.

They'll have to come to Azure. So, if it is any partner, and so so, fundamentally, you know so, uh, again, this is done in the spirit of what is it that we valued as part of our partnership, and we made sure while at the same time we were good partners to OpenAI, given all the flexibility they need.

So so, for example, Salesforce wants to integrate OpenAI, it's not through an API, they actually work together, train a model together, deploy it on, let's say, Amazon. Now, is that is that allowed or, uh, or do they have to No, for any Use it

Custom agreement like that, they will have to come run it. There there are some few exceptions to US government and so on that we made, but other than that, they'll have to come to Azure.

So, as Satya explained, as AI agents get more capable, you're gonna need more and more observability into what they're doing. You're gonna need to catch them when they're making mistakes, you're gonna need high-level summaries of what they're doing, and you're gonna need a picture of how everything that they're doing fits together. This is exactly what CodeRabbit provides. You just make a normal pull request, and CodeRabbit automatically reviews the PR. It generates a summary of changes so you can understand exactly what the PR's author was intending, and it uses the context from your full code base to provide line-by-line feedback on how things could be improved.

This is helpful whether you're reviewing a PR from a coworker or an agent.

In either case, CodeRabbit will write up its thoughts and flag any issues so that your teammate or your agent can go fix them. I've noticed that when I'm coding with agents, CodeRabbit catches a lot of mistakes that the models make by default. For example, the models have a bad habit of using old versions of libraries. So, in one session, I watched CodeRabbit cache a call to a an old model, figure out what the new version was, and then suggest that improvement.

Go to coderabbit.ai slash forkash to learn more. Uh, stepping back, a question I have is, you know, when we're t uh, w- w- walking back and forth through the, uh, the factory, one of the things you're talking about is, you know, Microsoft, you can think of it as a software business, but now it's really becoming an industrial business. Uh, there's all this capex, there's all this construction, and if you just look over the last two y- uh, two, um, two years, your sort of capex is, like, tripled, and maybe you extrapolate that forward, it just actually just becomes this huge industrial, uh, explosion.

Oh, their hyperscalers are taking loans, right? Meta has Meta's done 20 billion dollar loan at Louisiana, they've take they've done a corporate loan. It seems clear everyone's free cash flow is going to zero, um, which is which I'm sure Amie is, like, gonna beat you up for for even if you even try to do that. But, like, uh, what what what's happening?

I mean, I think, uh, I think the structural change, um, is what you're referencing, uh, which I think is massive, right? Which is I I describe it as we are now, uh, a capital-intensive business and a knowledge-intensive business. And, in fact, we have to use our knowledge to increase the ROI C on the capital spend, right? Because that's kind you know, look, at the hardware guys have done a great job, uh, of marketing the Moore's law, which I think is unbelievable, and it's great.

But if you even look, I think some of the stats I even did in my earnings call, which is for a given GPT family, right? Uh, the improvement software improvements of really throughput in terms of tokens per dollar per watt that we are able to get, uh, y- you know, quarter over quarter, year over year, is massive, uh, right? So, it's five x, ten x, maybe forty x in some of these cases, right? Just because, uh, how you can optimize.

That's s sort of knowledge intense intensity coming to bring out capital efficiency. So, that at at some level, uh, that's what we have to master. What does it mean? Like, somebody people ask me, what is the difference between, uh, you know, a classic old-time hoster and a hyperscaler? It was software. So, yes, it is capital-intensive, but as long as you have systems know-how, software capability to optimize by workload by fleet, that's why I think when when we say fungibility, there's so much software in it, it's just not about the fleet, right?

It's kind of the ability to evict a workload, eh, you know, and then schedule another workload. Can I, like, manage the that algorithm of scheduling around? Uh, that is the type of stuff that we have to be world-class at. And so, yes, so I think we'll still remain a software company, uh, but yes, this is a different business, um, and we're gonna manage. Look, at the at the end of the day, uh, the cash flow that Microsoft has allows us to have, um, both these arms firing on, you know, uh, well.

I it seems like in the short term, you have more sort of, um, credence on things taking a while, being more jagged. But in the maybe in the long term, you think, like, the people who say talk about AGI and ASI are correct, like, Sam Sam will be right, but eventually. Um, and I I have a broader question about what makes sense for a hyperscaler to do, given that you have to invest massively in this thing, which depreciates over five years.

So, s uh, so you if you have twenty, forty timelines to the kind of thing that somebody like Sam anticipates in three years, um, you know, what what is a reasonable thing for you to do in that world?

There needs to be an allocation, uh, to, well, call it research compute that needs to be done like you did R and D, right? So, that's the best way to even account for it, quite frankly. We should think of it as just R and D expense, and you should say, Hey, what's the research compute, and how do you want to scale it? And let's even say it's an order of magnitude scale, uh, in some period. Pick your thing.

Is it two years? Is it sixteen Months? What have you, right? So, that's sort of one piece, which is kind of that's kind of table stakes, that's R and D expenses. And the rest is all demand-driven, right? I mean, ultimately, you can, you'll have to build ahead of demand, but you better have a demand, uh, uh, plan, uh, that doesn't go completely off kilter.

So, do you buy? So, uh, these labs are now projecting revenues of 100 billion in twenty-seven, twenty-eight, uh, and they're projecting, you know, revenue keeps growing at this rate of, like, three x, two x See, there's Year.

A lot in the marketplace, right? There's all kinds of incentives right now, and and rightfully so, right? I mean, what y- what do you expect an independent lab that is sort of trying to raise money to do, right? They have to put some numbers out there such that they can actually go raise money so that they can pay their bills for compute and what have You? And it's and it's a good thing. I mean, someone's gonna take some risk and put it in there, and they've shown traction.

It's not like it's all risk without seeing the fact that they've been performing, whether it's open AI, whether it's Anthropic, so I feel great about the what they've done, uh, and we have Massive book of business with these chaps. So, therefore, uh, that's all good. But overall ultimately, there's two simple things. One is, you gotta allocate for R and D. You brought up even talent, you gotta s like, the talent for AI is at a premium.

You gotta spend there. You gotta spend on compute. So, in some sense, researcher to GPU ratios have to be high. Uh, that is sort of what it takes to be a leading R and D company in this world, uh, and that's something that needs to scale, um, and you have to have a balance sheet that allows you to scale that long before it's conventional wisdom and so on. So, that's kind of one thing. But the other is all about sort of knowing how to forecast.

A as we look across the world, right, America has dominated many tech stacks, right? Um, the US owns Windows, right, through Microsoft, which is the deployed even in China, right? That's the main operating system. Um, of course, there's Linux, which is open source, but, you know, Windows is deployed everywhere in China on personal computers. Um, you look at you look at Word, it's it's deployed everywhere. You look at all these various technologies, it's deployed everywhere. The thing that is quite unique and and and Microsoft and other companies have grown elsewhere, right?

They've built they're building data centers in Europe and in India and and and in all these other, you know, in Southeast Asia and in Latin America, Africa, right? All of these different places, you're building capacity. But this seems quite different, right? You know, to today, the political aspect of technology of compute, you know, you know, the the US. administration didn't care about the dot-com bubble, right? Um, it seems like the US. administration, as well as every other administration around the world, cares a lot about AI.

And the question is, you know, we're in a sort of a bipolar world, at least with US. and China, but Europe and and India and all these other countries are are saying, No, actually, we're gonna have sovereign AI as well. How does Microsoft navigate, you know, the difference of the nineties, where it's like, There's one country in the world that matters, right? It's America, and we do w- our companies sell everywhere, and therefore Microsoft benefits massively to a world where it is bipolar, where, hey, Microsoft can't just necessarily have the right to win all of Europe or India or, you know, Singapore, there's actually sovereign AI efforts.

What what is your thought process here, and how Do you think It's a great About this?

It's it's I think a a super, you know, critical, um, piece, which is um, I think the the key key priority for the US. tech sector and the US. government is to ensure that we not only do leading innovative work, but we also collectively build trust around the world on our tech stack. Right? Because I always say the United States is just an unbelievable place. It's just unique in history, right? It's 4% of the world's population, 25% of the GDP, and 50% of the market cap.

And I think you should think about those ratios and, uh, really, and reflect on it. That 50% happens because, quite frankly, the trust the world has in the United States, whether it's its capital markets or whether it's its technology and and its stewardship of what matters at any given time in terms of leading, uh, sector. So, if that is broken, uh, then that's not a good day for the United States. And so, if we start with that, which I think the you know, President Trump gets, the White House, David Sachs, everyone, uh, really, I think gets it, uh.

And so, therefore, I applaud anything that the United States government and the tech sector jointly does to quite frankly, for example, put our own capital at risk collectively as an industry in every part of the world, right? So, I would like, in fact, the USG to take credit for foreign direct investment by American companies all over the world, right? It's kinda like, uh, least talked about, but the best marketing that the United States should be doing is it's not just about all the foreign direct investment coming into the United States, but the most leading sector, which is these AI factories, are all being created all over the world by whom?

By America, uh, and American companies. And so, you start there, and then you even build other agreements around it, which are around their continuity, their legitimate sovereignty concerns around whether it's data residency, whether it's even what happens, um, uh, for them to have real agency and guarantees, uh, on privacy and so on. And so, the in fact, our European commitments, I think, are worth reading, right? So, we made a series of commitments to Europe on how we will really govern our hyperscale investment there, uh, such that really European, uh, Union and the European countries have sovereignty.

We're also building sovereign clouds in in France and in Germany. We have something called sovereign services on Azure, which literally give people key management services along with confidential computing, including confidential computing in GPUs, which we've done great innovative work with NVIDIA. Um, and so, I think I feel very, very good about being able to build both technically and through policy this trust in the American tech stack.

Hm. H- and how do you see this shaking out as you know, you do have this, uh, network effect with continual learning and things on the model level. Maybe you have equivalent things at the hyperscaler level as well. And, uh, do you expect that the country will say, Look, it's clear one model or a couple models are the best, and so we're gonna use them, but we're gonna have some laws around, well, the weights have to be hosted in our country, or do you expect that there will be, uh, this push to have it has to be a model trained in our country?

Maybe an analogy here is, like, people would l- you know, semiconductors is very important to the economy, and people would like to have their sort of sovereign semiconductors, but, like, TSMC is just better. And so semiconductors are so important to the economy that you will just go to Taiwan and buy the semiconductors you have to. Will it be like that with AI, or i is there

Ultimately, I think what matters is the use of AI in their economy to create economic value, right? I mean, that's the, uh, the diffusion theory, which is ultimately it's not the leading sector, but it's the ability to use the leading technology to create your own comparative advantage, right? So, that I think will fundamentally be the core driver. But that said, they will want continuity of that, right? So, in some sense, that's one of the reasons why I believe there's always going to be a check a little bit to sort of some of your points on, Hey, can this one model have all the runaway deployment?

It that's why open source is always gonna be there. There will be, by definition, uh, multiple models, that'll be one way, like, it's kind of the you know, uh, that's one way for people to sort of demand continuity and not have concentration risk is another way to say it is, right? Um, and So, you say, Hey, I want multiple models, and then I want an open source. So, I feel, uh, as long as that's there, every country will feel like, Okay, I don't have to worry about deploying the best model and broadly diffusing because I can always take, uh, what is my data and my liquidity and move it, uh, to another model, whether it's open source or on uh,

from another country or what have you. So, concentration risk, um, and sovereignty, right, which is really agency, those are the two things I think that'll drive the market structure.

The the thing about this is that this doesn't exist for semiconductors, right? You know, all refrigerators, cars have chips made In Taiwan.

Didn't exist until now until now, everybody's now like, you know, like, Even even There's

Then, right, America, you know, if Taiwan is cut off, there is there are no more cars, there are no more refrigerators. TSMC Arizona is not replacing any real fraction of the production, like, there it is it they're the sovereignty is a bit of, like, uh, a scam, if you will, right? I mean, it's it's worthwhile having, it's important to have it, but it's not a real it's not real sovereignty, right? And we're a global economy. We don't we

I think it's kinda like Dylan saying, Hey, at this point, we've not learned anything about sort of, uh, res- what resilience means and what one needs to do, right? So, it's kinda any nation state, including the United States, at this point, will do what it takes to be more self-sufficient on some of these critical supply chains. So, I as a multinational company, have to think about that as a first-class requirement, right? If I don't, then I'm not respecting what is in the sort of policy interests of that country, long term, right?

And I'm not saying they won't make practical decisions in the short term, right? Absolutely. I mean, the globalization can't just be rewound, right? I mean, all these capital investments cannot be made, uh, in in a way at the pace at which but at the same time, you have to kind of like, if I if you think about it, right? If somebody showed up in Washington and said, Hey, you know, you know what? We're not gonna build any semiconductor plants.

They're gonna be kicked out of the United States. Um, and and the same thing is going to be the true in every other country, too. Uh, and so, therefore, I think we have to, as companies, respect what the lessons learned are, um, you know, whether it's, you know, you could say the pandemic woke us up or whatever, but and nevertheless, people are saying, Look, globalization was fantastic, uh, it helped the supply chains be globalized and be super efficient, but there's such a thing called resilience, and we are happy you know, we want resilience, and so, therefore, that feature will get built.

At what pace, I think is the point you're making. It can't be like you can't snap your fingers and say all the TSMC plants now are all in Arizona, and with all of the capability. They're not going to be. But is there a plan? There will be a plan, and should we respect that? Absolutely. And so, I I feel that's the world. I wanna meet the world where it is and what it wants to do going forward as opposed to saying, Hey, we have a point of view that doesn't respect your view.

So, ju- just to make sure I understand, the idea here is each country will want some kind of data residency, privacy, et cetera, and Microsoft is especially privileged here because you have relationships with these countries, you have expertise in setting up these kinds of sovereign data centers, and therefore, Microsoft is uniquely fit for a world with, um, more sovereignty requirements.

Yeah, I would, I I I don't wanna sort of describe it as somehow we're uniquely privileged. Uh, I would just say I think of that as a business requirement

I see.

That we have been doing all the hard work all these decades And we planned to, and so my answer to Dylan's previous question was I take these re you know, whether it's in the United States, quite frankly, uh, uh, when you know, when the White House and the USG says, Hey, we want you to allocate more of your, I don't know, wafer starts to, uh, uh, fabs in the US, we take that seriously. Uh, or whether it is data center and the EU boundary, we take that seriously.

So, to me, um, respecting what I think are legitimate reasons why countries care about sovereignty and building For it as a software and a physical plant is what I I would say we are gonna do.

And as we go to, like, the bipolar world, right, US, China, Um, there Is there is a Lot around You know, American tech does not you know, it's not just you versus Amazon, um, or you versus, you know, Anthropic or you versus Google. There Is a whole host of competitive com competition. How does how does America rebuild a trust? What do you do to rebuild the trust to say, actually, no American companies will be the main provider for you, um, and how do you think about competition with up and coming Chinese companies, whether it be, you know, ByteDance and Alibaba or DeepSeek and Moonshot?

And this would just add to the question. One concern is, look, we're talking about how AI is becoming this sort of industrial capex race, uh, where you're just rapidly having to build quickly across all those supply chain. When you hear that, at least up until now, you just think about China, right? This is, like, their comparative advantage, and especially if we're not gonna moonshot to ASI next year, but we it's gonna be this decades of build outs and infrastructure and so forth.

How do you deal with Chinese competition? Are they privileged in that world?

Yeah. So, it's a great que- I mean, in fact, you just made the point of why I think trust in American tech is probably the most important feature. It's not even the model capability, maybe. It is, like, can I trust you, the company? Can I trust you, your country, and its institutions, to be a long-term supplier, may be the thing that wins the world.

I think that's a good note to end on. Satya, thank you for doing this.

Thank you so much. Thank you. It's thanks. It's such a pleasure. Such a pleasure. It's awesome. It's like, man, you two guys are, like, quite the team.

Hey everybody, I hope you enjoyed that episode. If you did, the most helpful thing you can do is just share it with other people who you think might enjoy it. It's also helpful if you leave a rating or a comment on whatever platform you're listening on. If you're interested in sponsoring the podcast, you can reach out at twarcash dot com slash advertise. Otherwise, I'll see you on the next one.