# Satya Nadella on Open vs Closed Models: Who Wins | Why Companies Need to Own Their Own Intelligence 20VC with Harry Stebbings · 2026-10-26 Satya Nadella, CEO of Microsoft, argues AI models should be assumed compromised and contained in a Zero Trust architecture because alignment no longer falls out of scaling. He calls chain of thought unfaithful, wants tamper-proof agent logs via Agent 365, and says transparency demands could favor open-weight models. He argues every firm must own its intelligence — data, traces and private evals as IP, citing his 'reverse information paradox' — and sees a new implementation economy emerging. He doubts full-stack vertical integration wins, expects commodity tokens to dominate traffic, and says $100 billion training runs require world GDP to inflect, pointing to NHS Copilot and elastic software demand. In quick-fire, he reveals his changed mind: alignment will not just fall out of scaling. ## Questions this episode answers ### Why does Satya Nadella say we should assume AI models are compromised? Satya Nadella says scaling laws are working and models are becoming more capable, but his old assumption that alignment would just "fall out" of scaling no longer holds. Since mechanistic understanding is elusive, he argues companies should treat models with a Zero Trust architecture that can contain any malfunctioning agent before it does harm. [1:24](https://20vc.podhood.com/7fd44fa7-0eb4-4c50-8d23-2f45adb244cd?t=84000) ### What share of AI traffic will go through open versus closed models? Satya Nadella tells Harry Stebbings it depends on the frontier: most capabilities saturate, and once saturated, Open-Weights or cheaper closed models suffice, especially if you can train on your own traces. He predicts there will always be a market for frontier models, but predominantly token usage will be commodity token usage. [7:05](https://20vc.podhood.com/7fd44fa7-0eb4-4c50-8d23-2f45adb244cd?t=425000) ### Why does Satya Nadella think every company will have a private eval benchmark? Satya Nadella says every firm will have a private eval benchmark as its core IP, with models hill climbing on it. He compares it to early PC adopters in the late '80s and early '90s: what seems rarefied today, like being an AI-native company, will be passé in about five years, just as even 500-year-old Lloyd's became cloud-native. [10:44](https://20vc.podhood.com/7fd44fa7-0eb4-4c50-8d23-2f45adb244cd?t=644000) ### Can the world support $100 billion AI training runs? Satya Nadella says a training run is ultimately just a percentage of R&D spend, so it requires enough revenue, which comes from contributing to world GDP growth. He notes supply-side constraints like power and silicon shortages are long lead items, and says the bigger question is what the world GDP growth rate will be. [16:16](https://20vc.podhood.com/7fd44fa7-0eb4-4c50-8d23-2f45adb244cd?t=976000) ## Key moments - **[0:00] Intro** - **[1:24] Zero trust** - [1:24] Satya Nadella says companies should assume AI models are compromised and run them inside a Zero Trust architecture - [2:44] "I used to always think that alignment will just fall out of scaling — I no longer believe that": Satya Nadella on why scaling broke his alignment assumption - **[4:46] Chain of thought** - [4:46] Satya Nadella: chain of thought is unreliable because models can be unfaithful, and hiding it makes the problem worse - **[5:42] Open vs closed** - [5:42] Satya Nadella on Chinese open-weight models: unaligned AI is a danger to China too, and Microsoft's MAI models will ship as licensed weights - [7:05] Satya Nadella predicts most token usage becomes commodity within 24 months, with a permanent market only at the frontier - **[9:07] Own intelligence** - [9:07] Satya Nadella's 'reverse information paradox': AI buyers now need trade-secret protection because thumbs-up data leaks value to model companies - [10:44] "Every firm will have a private eval benchmark, which is their core IP": Satya Nadella on companies hill-climbing models on their own evals - **[11:09] Evals everywhere** - **[12:17] Implementation economy** - [12:17] Satya Nadella's Unilever story: an agentic order-to-cash system that swapped to a cheaper MAI model by training on its own traces - **[14:11] Vertical integration** - [14:16] Satya Nadella doubts one vertically integrated AI stack wins: 'every day you wake up and there's more competition in every layer' - **[16:01] $100B training runs** - [16:01] Satya Nadella on $100 billion training runs: they only make sense as a share of R&D if world GDP actually inflects - **[17:55] GDP inflection** - [17:55] Satya Nadella says AI can break the world economy's 2% growth ceiling, starting with healthcare deployments like Copilot in the NHS - **[19:10] Beyond coding** - **[20:45] Quick-fire** - [21:28] Satya Nadella on changing his mind: 'I was one of the first believers in scaling laws — alignment may be a harder, different problem' - [22:36] Satya Nadella's dream for AI: convergence growth that lifts the Global South, echoing his father's work as a development economist ## Speakers - **Harry Stebbings** (host) - **Satya Nadella** (guest) ## Topics Tech Market Trends ## Mentioned Microsoft (company), NHS (company), Rolls-Royce (company), Unilever (company), Agent (product), Copilot (product), GitHub (product), MAI (product) ## Transcript ### Intro **Harry Stebbings** [0:01] And Dario, and Ilia, and Sam, and Greg, and Elon were all, I guess, one happy family. They were really the believers in scaling laws, and that's what I believed in back in 2018, 2019. I think the biggest change now I have is: I used to always think that alignment will just "fall out." I no longer believe that. Now, this show, honestly, is quite an emotional one for me. I started 20VC from my bedroom in London, I was 18 years old, I had no money, I had no contacts in my phonebook, and today I have the chance to sit down with Satya Nadella, CEO of Microsoft, one of the greatest companies of the last 3 or 4 decades. And I will just say, on Satya, the warmth that he was able to imbue in this interview in such a short space of time was almost unlike anything I've seen before. The clarity of his arguments and his ideas, his ability to go very deep very quickly, it was almost unlike anything I've ever seen before. It was a show I'll always remember, and I really hope you enjoy it. Ready to go? Satya, thank you so much for doing this with us. I would say, you know, thank you for coming to the studio, but you're actually hosting us here, so really appreciate you taking the time. **Satya Nadella** [1:19] No, like, thank you so much for having me on the show, Harry. It's been a long time coming, but I'm finally glad we were able to do it. **Harry Stebbings** [1:24] Now, you wrote a seminal piece 10 days ago or so, and you said that we should assume a model is compromised, and I just wanted to expand on that. Does that mean that we can't or shouldn't rely on the disclosures of the model provider companies? ### Zero trust **Satya Nadella** [1:39] Yeah, I think the challenge, the way I see itright now, is the models are becoming more and more capable,right? Scaling laws are working. Except one of the assumptions I had was that you'll have alignment just fall out of scaling, and it doesn't seem to be. And there's more science challenges there that need to be tackled,right? And mechanistic understanding is also elusive. And so given that, I think the best posture to have if you want to continue to use these sophisticated models is to treat these models in a trust architecture. I mean, after all, Zero Trust is not something new,right? We've had Zero Trust architectural approaches. So the claim, at least I'm making, is: let's create an engineered Zero Trust environment that can contain any malfunctioning agent before it does any harm,right? So that's the point. Because if you really want to have agents that are more capable and yet less aligned, then you need to have a trust system that you can rely on. **Harry Stebbings** [2:39] Can I ask, what does that trust system look like? Who's the architecture of it? Who's the arbiter of it? **Satya Nadella** [2:44] It's a great point. And so that's why I kind of look at it and say, you leave AI safety outside of this,right? Because that alignment, for example, at least my take is, that it really mostly has to show up in training,right? In fact, maybe even in pre-training. I don't think you can even RL your way to alignment. You have to truly get alignment to be something that truly gets embedded in the way it'sright at the very beginning of it. And so that's the model itself, hopefully at some point, will get aligned. So if you leave AI safety, then the trust architecture is going to be just, you know, what a company like Microsoft does, or all the cyber folks do. So what we have done, for example, is we built these two things,right? Agent 365 and Perception,right? What does it do? It gives every agent an identity. It then logs all of its activity, all its tool calls, in fact, every model call and the COT in it. In fact, that's where I'm also asking that all the COT better be in English and not in New Relief. And then you can check using multiple models. It better be tamper-proof,right? You can't have the agent go modify its own logs. And then on that, you set policy, and then you behaviorally monitor the activity of the agents so they're confirmed to the policy. So that's the architecture. **Harry Stebbings** [4:02] We don't have that architecture today. Is that just symbolic of where we are in the progression? **Satya Nadella** [4:07] No, that architecture exists. So there are two, I think people conflate what I'll call an engineered security Zero Trust architectural system for containment and AI safety,right? Because I think people say, well, AI safety is hard and elusive and not yet achieved, so therefore there must be nothing we can do on security. And I'm not saying that these two things are all got to come together at the same time. In fact, I would just say, hey, we have to deal with the situation as it exists. So let's build a trust system outside, in fact, of the model, and then just be super aggressive in setting policy and monitoring for policy deviation. ### Chain of thought **Harry Stebbings** [4:46] On the monitoring side, if we think about chain of thought, how reliable is that? **Satya Nadella** [4:52] It's not,right? We know that the chain of thought itself can be unfaithful, and in fact, it can be made worse when the model company says, well, it's unfaithful and I'm going to hide it. And so it becomes even more problematic. And that's where I think a little bit of the push for standardization and the COT, at least for trusted partners being available as is, so that it's inspectable, it's auditable, and I can check using multiple models and what have you,right? So that's going to be one of the, in fact, that's one of the things that may even drive more Open-Weight, quite frankly. Because if you can't trust, it's funny, closed-fed models are more powerful, but if they kind of don't recognize the need to be more transparent, especially as their trusted partners use it, then we may have a problem. **Harry Stebbings** [5:42] Can I ask you, when we think about that rise of open source, I completely agree with you in terms of closed being more powerful, open source today is predominantly Chinese models. Is that a concern then when we think about safety being baked into pre-training, when we think about disclosures, as we said there, and then juxtapose with the majority of open source being Chinese? ### Open vs closed **Satya Nadella** [6:01] Yeah, I mean, I think there are two sets of things. I do believe, if you sort of step back, somehow AI safety not being there is not a good thing for China, it's not a good thing for the United States, it's not a good thing for the UK. So I think that there's a part of everyone, because this is not something that somehow an unaligned Open-Weight Chinese model will probably attack even China. So it's not like you can sell the rest, think that this will only be used as an attack vector. So setting that aside, I'm pretty pleased with some of the progress even the Open-Weight models in the West are making. And I'm, you know, obviously, you know, Jensen's working on it, the other folks are working on it. We will have even the MAI models that will be not Open-Weight, we'll talk, I think of it more as licensed weights, think of it that way. I want to turn this into just a simple, straightforward software business as opposed to a network effects intelligence business. So I think there will be different business models. And I ultimately trust in your ability to use those models and control those models is what's going to determine what gets used. **Harry Stebbings** [7:05] As a venture investor, I have to predict where traffic will go today between closed, open. When you think about the two different dimensions, which is like traffic and dollars, if we look forward, say, 24 months, what percent of traffic goes through open versus closed, do you think? **Satya Nadella** [7:21] It's a great question. You know, I think it depends on what the frontier looks like,right? Which is, I think most things seem to get saturated. And when it gets saturated, Open-Weights or even just cheaper models of even closed weights seems to work. Because especially one of the things is, if you can train on your own traces, then you really don't need the latest model to be able to solve that same problem if the problem has not changed,right? So if you have, you know, if you have an eval that you want to continuously keep moving forward, then you may want to pay. But even there, by the way, I'm very thrilled about this hydrofusion stuff we now have in GitHub, where, you know. **Harry Stebbings** [8:04] I just saw this. **Satya Nadella** [8:05] Yeah, because the cool thing there is it's a learned model. It knows, it's anyway, we are now going into these agent swarms and multiple sub-agents. So I can call a lot of sub-agents that even happen to be local models. And so I think my suspicion is that there will always be a market for frontier. There will always be some new things. Science may be the next thing that, you know, frontier is being set at. But I think predominantly token usage will be commodity token usage. **Harry Stebbings** [8:35] Does that mean it's open source predominantly? **Satya Nadella** [8:37] It may be, it may be Open-Weights, or as I said, it can also be cheap closed weights,right? Because there will be lots of price competition, let me put it that way,right? Because if you look at my own usage inside of Copilot, we use a lot of closed source, cheaper models along with Open-Weight models. **Harry Stebbings** [8:56] You said about kind of the importance of owning your own intelligence for some of your amazing customers that you have. You were talking about some in the UK. Is that the model environment that we have moving forward? Specialized intelligence for companies with their own data? ### Own intelligence **Satya Nadella** [9:07] Yeah, so it's, I think this is, to me, one of the most important things,right? Which is, what's the future of a firm in an era of intelligence,right? Because if you go back even to Hayek, his fundamental description of an economy is local intelligence. Tacit knowledge is distributed throughout the economy,right? Without that, you know, there can't be only one firm and it can't be one intelligence. And so that's sort of the foundational thing. In fact, capitalism works because of that. Price discovery works because of that. Or pricing mechanism works because of that. So therefore, I think it'll be important for every firm to be able to accumulate knowledge at increasing scales using AI versus leak. That's why I called it the reverse information paradox,right? Which is, you know, Kenneth Arrow kind of described this paradox of information where if you're a seller and you disclose the secret without a patent or a trade secret protection or copyright protection, you leak the value. Now, for the first time, the buyer needs the protection. Because the buyer, even if, you know, most people think of it as, oh, my data. It's even my exhaust of using it,right? Thumbs up, thumbs down signal. Like, you're creating value for the model company versus capturing that value for yourself. So I think one of the things that at least I ascribe is that we need to have, for this to be stable, more of that surplus that gets created by use of intelligence to remain at the firm. And by the way, it doesn't need to be only by embedding your knowledge in weights. It can be in context. It can be in your evals,right? You may have a benchmark. In fact, one of the things that I'm very gung-ho about is we all become, every firm will have a private eval benchmark, which is their core IP. And every model comes in and tries to sort of hill climb on that eval. And that, to me, is how a firm works. ### Evals everywhere **Harry Stebbings** [11:09] I love this. Private evals, benchmarks for each company. This is for a rarefied era of companies that would seem your Harveys or your Liguras or the kind of very forward-thinking AI companies. Will that disperse to every company in the world you mentioned? **Satya Nadella** [11:25] Yeah, it happens. Yeah, it happens, Harry, always,right? Which is, you know, there was a day when I remember in the late '80s, in the early '90s, that the few companies that had deployed PCs and were doing forecasting, which felt magical, they were able to put it in Excel spreadsheets, send it in email, and avoid the fax machine and inter-office memos. And it felt like, oh, wow, when will this happen? And it took 10 years maybe, and it spread out throughout. This time around, what is considered, oh, I'm an AI native company and I do this. I think, you know, five years from now, it'll be pretty passé, I would imagine. And then that's kind of what, I mean, these tech cycles always start like that,right? Which is even cloud. I remember in the beginning of cloud, it was like the big deal was, oh, who's cloud-native and who's not? And, you know, if you're not cloud-native, and now everybody is,right? Even Lloyd's, you know, which is what, 500 years old, is cloud-native. And so it happens. ### Implementation economy **Harry Stebbings** [12:17] Do we move into an implementation economy then? Because when we think about these customers, as you said, some of your very big customers, they will need heavy FDE support, much more so than they have done in traditional years when the last 10 years. Will we have a very heavy implementation economy to transition existing customers into this? **Satya Nadella** [12:35] Yeah, it's interesting,right? The implementation economy, like systems integrators have always existed,right? You know, we've depended on implementation economy for lots of other technology diffusion. So this one, I think, has different contours,right? Because it's the combination of tech and business, if I could say. But it's not classic management consulting, it's not classic systems integration, and it's not classic sort of technology support. It's sort of a, like, for example, think about this,right? You want to start, like when I look at my FDE engagements,right? Even take the Unilever example I talked about at the keynote today. It's fascinating. They've built an agentic system to say, hey, we're going to really have order to cash automated using an agentic system, plus their trade claims. And by the way, they start using a frontier model, and then they substitute for, in fact, I think they're using currently an MAI model for all the trade claims because it's sort of a lower cost. And it's, by the way, they control it,right? Because they're able to train on their traces. And so we had some FDEs, but a lot of who got deployed, and now their IT team itself is, you know, managing it all. But it started with what's the business, what's the eval, what's the way to conceptualize the agentic architecture? So it's sort of business consulting combined with technical consulting. And I think, yeah, you'reright in saying that there will be a new set of skills that will get built in the ecosystem. Today, I think all of us have, quote unquote, our own FDEs, but FDEs will become like, you know, the traditional systems integration will become new FDE. ### Vertical integration **Harry Stebbings** [14:11] You've mentioned, obviously, MIA a lot, MAI, MIA. **Satya Nadella** [14:16] Yeah. **Harry Stebbings** [14:16] Gosh. But when we look at a lot of different players, it's kind of this realization that you need to be full stack, from chips to models to applications. Is that the future? You have to have vertical ownership. And what does that mean? **Satya Nadella** [14:30] Yeah, it all depends on the industry structure,right? You know, I think that I would say today, you know, there'll be, who knows, five, six Open-Weight, you know, closed models all competing, maybe five, six, you know, Open-Weight models all competing. And so therefore, at some level, if you're building an application like the agentic system, you could even have something like what I said, which is a router with even a decision engine built into it, and then you just route to the best model for a given task. So that's a possibility. Of course, there's always the risk that there could be one vertically integrated model and app, app tier, and all the way to the silicon and system. It's unlikely to me,right? I thought at some point that it might be the case that there will be one. If anything, it's going the other way, which is every day you wake up and you say, wow, there's more competition in every layer of the stack, whether it's in silicon, whether it's in systems to manage the silicon, in the model layer, in the app layer. And so I think that's a healthy sort of direction of travel, but you never know,right? Because this is one of those things where it all depends on what becomes the dominant design. But my hope is that this becomes a little bit of a horizontal specialization with multiple winners in each layer, and then the ability for people to compose. **Harry Stebbings** [15:53] The thing I didn't see is the depth of each layer. Like what I know now that I wish I'd known 12 months ago, it's just like the infinite demand for inference. **Satya Nadella** [15:59] Yeah, that's a great point. **Harry Stebbings** [16:01] Yeah. And I wish I'd known it 12 months ago, I would have done some very different investing. Speaking of that, we've had like Mustafa Suleyman saying training rounds of $100 billion are coming, and only a handful of labs can pay for that. ### $100B training runs **Harry Stebbings** [16:12] $100 billion training rounds. Do we have the energy, infrastructure, and power to enable this scale across the stack, but especially when we do? **Satya Nadella** [16:23] Yeah, I mean, the way I come at it is, ultimately, it has to come from the world GDP inflecting, and then that allowing someone to then have as a percentage of R&D,right? Because when somebody says training run, it's just a percentage of R&D,right? Sort of companies that are heavy R&D spend what, 15, 20%, you know, on R&D. And so therefore, yeah, at some point, if you have enough revenue, you can spend enough on training. And then, of course, you'll have enough revenue if you're contributing to the world economy growing so that you can get your fair share, so to speak. And then you also bring up the other side, which is the supply side, which is we have power shortages, we have silicon shortages, we have, you know, all kinds of things today. They will all, they're long lead items,right? You just can't, you know, snap your finger and have power generation, or you can't have a new fab. And so now, interestingly enough, given the demand, everyone's building. And so the question is, when will all of this come online? What will be the demand supply? No one knows,right? Quite frankly, any one of us who are sitting here today will just be speculating. But my feeling is that if you start by thinking of the stable state would be just like today,right? The income statements are suddenly not going to be different,right? For a company, it's going to be, hey, I have some revenue, I have some cost of revenue, and I have to have some profit. And if that is going to be true, then the question is, what is the world GDP growth rate? That's the bigger thing for me. **Harry Stebbings** [17:55] In terms of the inflation of GDP growth rate, the world has kind of stabilized at 2% a year for a very large amount of time. Does AI have the ability to seriously inflect that in a short amount of time? ### GDP inflection **Satya Nadella** [18:06] I think so, and I somewhat hope so, because that is the point,right? When I look at, let's take even as a percentage of any of our GDPs, the amount of money we spend in our healthcare,right? So that's why I look at even the deployment of Copilot in NHS and say, wow, if it can improve patient outcomes and reduce cost in the healthcare, that is by definition, you know, helping the GDP of the United Kingdom. And so I think that there are these pockets of the economy which are ripe for adoption of AI and really, in fact, there's the least amount of displacement, the most amount of benefit. And so that, I think, is where I think we'll have to find the best fit. Software development is another one,right? Which is, it's fascinating to see how much the demand is elastic for software development,right? Which is, that means the IT backlogs were huge, and none of us even recognized it. ### Beyond coding **Harry Stebbings** [19:10] I was taught that you had to be the top one or two in a market to win in venture. We're in, you know, factory, which is fifth, and it's winning. Fifth. **Satya Nadella** [19:19] Yeah. **Harry Stebbings** [19:19] This breaks all rules of venture. The question that I have, again, slightly naughty note on the schedule, but it's just like, will we see the tipping in terms of capability advancement in alternative functions, finance, sales, marketing, that we've seen in coding? Because coding's been the mother lode. **Satya Nadella** [19:34] I think that that's a good point. You know, people talk a lot about, you know, do you have a verifiable reward? And, you know, it's easy to do these, because a lot of the magic has been RL on top of these pre-trained models. That is what, you know, has helped coding. Can we crack that in knowledge work? Seems much harder. But again, all the examples I talked about, sort of, you know, in contact centers, in these supply chain systems, in these marketing slash or cash to order systems, I think that it's possible, and we're seeing it. Is the demand as elastic as coding? I think it may be science, maybe as elastic,right? Or engineering in general. Like, by the way, you know, if I look at even what's happening inside, not just in software development, let's take silicon development. That's another thing that's going through a radical shift,right? Your ability to tape out much faster, simulate everything in silico is much better. Your first spin can completely come out. **Harry Stebbings** [20:32] But even just like optimization of maintenance reports, like you said with Rolls-Royce, which I would never. **Satya Nadella** [20:36] It's unbelievable. Like, I mean, it's a thing that they've been dealing with forever, just schedule maintenance. And I did not realize that, wow, that's a big optimization problem that you can go now run. ### Quick-fire **Harry Stebbings** [20:45] So I've watched many, many of your shows, and we're going to do a quick-fire round. **Satya Nadella** [20:49] Yeah. **Harry Stebbings** [20:50] And people haven't asked some of these questions, which I've always wanted to know, and I think everyone wants to know, but, you know, people haven't asked them. What's the exercise routine given the travel schedule, Satya? **Satya Nadella** [21:00] Well, I mean. **Harry Stebbings** [21:01] You're in good shape. **Satya Nadella** [21:02] Well, I try to be. I don't know whether it's all well hidden. But, you know, the one thing I've gotten into the habit, you know, this has been now maybe 25 years even on, which is every morning I get up and, you know, I used to be able to run. Unfortunately, my knees have gone away. And so now at least I walk with incline. Even this morning, I was able to get a half hour in. And so that's what, in spite of any jet lag, wherever I am, I get hit the gym every day. **Harry Stebbings** [21:28] Love it. I too have bad knees. It's my age as well. What have you changed your mind on in the last 12 months? **Satya Nadella** [21:35] Last 12? It might be, because after all, I was one of the first, yeah, obviously when Dario and Ilya and Sam and Greg and Elon were all, I guess, one happy family, they were really the believers in scaling laws. And that's what I believed in back in 2018, 2019. And I think the biggest change now I have is I used to always think that, you know, alignment will just fall out. I no longer believe that. So I think this alignment thing may be a harder, different problem. And so, yeah, that's the one thing I've changed my mind on. **Harry Stebbings** [22:11] What looks crazy today, but will be incredibly common and accepted in five years? **Satya Nadella** [22:15] You know, it might be that thing that we just talked about, these evals,right? What is a firm? I think a firm would be an AI benchmark. **Harry Stebbings** [22:25] Final one for you. There are many things to be excited about over the next 10 years. For you specifically, like for me, it's my mother's got MS. I'm so excited for the drug development that could be capable. What are you most excited for? **Satya Nadella** [22:37] I mean, that's a great one, I think. One of the things that I've always been, I would say, been looking forward to is a general purpose technology, even as powerful as this. And people say, oh, this is not just another general purpose technology, whatever this is. I hope this spreads, diffuses throughout the world faster than anything else, and throughout all sectors of the economy, unlike anything else. And truly, for once, there is a real shift in the frontier of that global growth that happens where it's not about just counting winners and losers, but it's kind of like everyone gets lifted. Like that's been my dream. In fact, my father was a development economist, and, you know, it was always about convergence growth. When will the Global South get something that just suddenly boosts that? I hope this is it. **Harry Stebbings** [23:34] Satya, this has been such a joy to do. Thank you so much for taking the time. **Satya Nadella** [23:37] Absolutely. It's a pleasure, Harry. --- This library is powered by PodHood (https://podhood.com), the podcast website platform.