February 27, 2026
WHEN: Today, Friday, February 27, 2026
WHERE: CNBC’s “Squawk Box”
Following is the unofficial transcript of a FIRST ON CNBC interview with Amazon CEO Andy Jassy and OpenAI CEO Sam Altman with Andrew Ross Sorkin on CNBC’s “Squawk Box” (M-F, 6AM-9AM ET) today, Friday, February 27. Following is the video from CNBC.com: https://www.cnbc.com/video/2026/02/27/a-very-strong-long-term-partnership-openai-ceo-and-amazon-ceo-on-new-strategic-partnership.html.
All references must be sourced to CNBC.
ANDREW ROSS SORKIN: Welcome back to “Squawk Box.” A massive deal this morning. OpenAI now raising $110 billion in new funding. It values the company at $730 billion. Investors in the round include Softbank, Nvidia, and importantly, the biggest one now, Amazon, committing $50 billion in funding for this round. And the pair now announcing a new multi-year strategic partnership, which may reshuffle the landscape of AI all over again. Joining us now for a first on CNBC interview is Amazon CEO Andy Jassy and OpenAI CEO Sam Altman. Good morning to both of you on what appears to be a watershed transaction. Sam, I'll start with you, $110 billion, $730 billion valuation, but explain, if you could, how you think this deal shifts the balance. I was suggesting it shifts the balance in large part because the relationship you've had for so very long was with Microsoft, and now you have Amazon as well, and in certain cases, on an exclusive basis, for a big part of the – sort of the next push.
SAM ALTMAN: Good morning. Thanks for having us. We're super excited about this deal. We will continue to have a great relationship with Microsoft. We're excited to have a great relationship with Amazon. AI is going to happen everywhere. It's transforming the whole economy, and the world needs a lot of collect – collective computing power to meet the demand. The world also needs a lot of new kinds of products. And we'll continue to do our work with Microsoft, but we're very excited about the new product work we'll do together with Amazon. And I think, as you said, it will be a great, very long-term partnership.
SORKIN: We're going to get into all of it. Andy, $50 billion a huge investment from Amazon. As part of that investment, there's a whole bunch of component parts around the partnership that you're going to have with this company. But interestingly, in the press release, you talk about $15 billion up front, $35 billion that will come after, if certain conditions are met. And I think there's going to be a lot of people are going to read that and say, what does that mean?
ANDY JASSY: Well, you know, thanks for having me. I appreciate it, and we're excited about what we announced. And yeah, we have a significant investment that we're making. It's in two tranches, as we mentioned. And you know, the first tranche will be the next month here, and $15 billion and the second tranche will come when certain milestones are met. And we're excited about the investment we're making. If you think about it, it's so early right now in the AI space, and OpenAI is off to an amazing start. They're going to be one of the very big winners, we believe, long term. I think we can help them quite a bit as part of this partnership. I think just having their offerings available in bedrock, you know, being able to leverage training, which gives them 30 to 40% better price performance matters a lot in AI, and I think it's going to be a very strong long-term partnership where we're going to also be happy about the investment. It'll yield a good return for Amazon over a long period of time.
SORKIN: So Andy, one of the things that's interesting about this is OpenAI it says will “consume 2 gigawatts of Trainium capacity through AWS infrastructure to support the demand” for what's called “Stateful Runtime Environment, Frontier, and other advanced workloads,” some of which is going to be exclusive to Amazon. My question to you is twofold. One is, how long is that exclusivity period for? And two, how much revenue do you imagine you will ultimately get paid by OpenAI?
JASSY: Well, you know, there are multiple components of this. If you think about, you know, we've collaborated and invented together on a Stateful Runtime Environment powered by OpenAI's GPT models that will be available in bedrock. And if you think about, if you're an AI application developer, you don't want to start from scratch every time you're actually using models. And so being able to access state, whether it's memory or identity, or being able to call tools or call out to compute, being able to do that in a Stateful way, where we're together, training those – the Stateful Runtime Environment on AWS’s infrastructure, there's nothing else like that today. It's really the next generation of how AI developers are going to build their AI applications. And then, you know, the training today, we now have the two largest AI labs who are both significantly betting on Trainium, which is exciting. And, you know, and the relationship we're building is very unique, and it's long term, and we'll have this partnership for a long time.
SORKIN: Sam, can you speak to that in terms of just what you think the business ultimately looks like with Amazon? Obviously, they just made a big $200 billion infrastructure investment. A lot of people had asked questions about it. Now this may very well be part of that answer.
ALTMAN: Yeah, we continue to see just rapidly growing demand. The Codex in particular, I think it grew like 30 something percent in the last week. And that is an indication of what's happening with enterprise and what people are ready to do now. As we started to look at the growth that is coming, we realized we did not have enough compute to support that. We're excited, thrilled to get this additional Trainium capacity, and I think it'll go to great use with customers. We also think, as Andy was saying, that Amazon can deliver so much more to us in terms of new demand and new opportunities in the market, that there will be huge incremental revenue here into the whole ecosystem. And we're thrilled to get to do that.
SORKIN: One question which I imagine investors will ask, which relates to maybe the question I asked Andy about the conditions on the 50 billion. It sounds like it comes in tranches, Sam, 15 now, 35 later, with some kind of condition. And I just wonder whether that is similar in terms of the terms for every other investor in this round. There's been lots of speculation that one of the things that every investor in these new rounds wants, of course, is eventual IPO.
ALTMAN: We are open to going public at the right time. There are advantages to being private. There are clear advantages to being public. In addition to all of the reasons that we get more access to capital, I also think that if these AI companies are as important as we think they may end up being, we want all investors to have access to them. So yes, there is definitely some excitement from various investors of ours to go public, and it is something that we are open to doing, but we need to figure out the right time to do it.
SORKIN: Sam, help me with this, though, because you know that people oftentimes look at some of these transactions as what they call circular. They look and say, a vendor, effectively is going to make an investment in your company, and I – and on the other end, you are going to be making effectively an investment or buy services from that company. How do you think we should think about that?
ALTMAN: I get where the concern comes from, but I don't think it matches my understanding of how this all works. This only makes sense if new revenue flows into the whole AI ecosystem. If people are not willing to pay for the services that we and others offer, if there's not new economic value being committed, then the whole thing doesn't work. And it would just it would be circular. But revenue for us, for other companies in the industry, is growing extremely quickly, and that's how the whole thing works. Now, given the huge amounts of money that have to go into building out this infrastructure ahead of the revenue, there are various things where people, finance chips invest in each other's companies and all of that, but that is like a financial engineering part of this and the whole thing relies on us going off – or other people going off and selling these products and services. So as long as the revenue keeps growing, which it looks like it is – I mean, demand is just a huge part of my day is figuring out how we're going to get more capacity and how we're allocating the capacity we have. Then, I don't think it looks circular, even though the need to finance this, given the huge amounts of money involved, does require a lot of parties to do deals together.
SORKIN: Andy, how do you think about that?
JASSY: I see them as different. You know, I think that, you know, OpenAI wouldn't use Trainium if it wasn't compelling for them. You know, the fact of being able to save 30 to 40% in price performance, that's a big deal when you do AI at the scale that OpenAI does. And I would also say that for us, the investment is because of what I said earlier, which is, you know, it's – this is the most significant transformation in technology, maybe business in our lifetimes. And we're incredibly – even though it's growing really quickly, we're incredibly early in this and OpenAI is going to be one of the really big winners here, and I think we can help them be more successful, as I mentioned earlier, and we're going to like that investment over a long period of time.
SORKIN: Andy, you know, you had announced earlier this year that you're going to be spending $200 billion – there were a lot of folks in the investment community that said, oh my goodness, that's a lot of money. And I suggested just now that maybe this was part of the answer. Did you have line of sight, if you will, that this was a deal like this was possible at that point?
JASSY: Yes. Sam and I have been talking and our teams for a while, and that was very much in our projections at that time. And you know, as I said during the call, when we said we were going to spend about $200 billion in capex this year, we have very strong demand signals, very clear line of sight to those. This is a good example of that. Remember, the capital we spend in ‘26 is for infrastructure that will be put in place, you know, 18 to 24 months later. And so, we have very high confidence that we're going to be able to monetize what we're spending, and we're going to like the economics a lot in that business.
SORKIN: And when you think about that investment, and maybe both of you can speak to this, do you think there's genuinely just a massive upfront cost, and at some point that the cost really does tail off, or from where you sit today – I mean, this is what I think some people are concerned about, is that it's just an exponential cost. That it's just going to be year after year after year, and there's not going to be a moment where the investment stops, if you will.
JASSY: I'll start. You know, the way the AWS business works in such a way where the faster we grow, the more capital we put out there. And because the model works, you know, we're investing in capital in data centers, power, land, chips, hardware, networking gear. So, all this infrastructure 18 to 24 months in advance of when we can monetize it. And then we spend the money in the first year, but a lot of these assets that we're investing in are 30 to 40 year useful life assets like data center. And so what happens is, you spend the money in year one, but then you actually monetize over a long period of time, and you really like the return on investment in capital over time. You saw that in the early days of AWS, you know, as we are growing like crazy, just on the CPU side, where we were investing a lot of capital and the free cash flow looks different than it does once the growth rates start to level off. And so, as the growth rates are really high, we'll spend more capital, but we like the operating income and the return on invested income over a long period of time, because those assets we can monetize over a long period of time.
SORKIN: Sam, how do you think about that? And both on the capex side for some of the folks providing these data centers, but also for your own business.
ALTMAN: Yeah, I very much agree with what Andy said. The investment needs to happen early, but the revenue growth and the ability to monetize that investment looks like it will continue to be steep over a long period of time. And like Andy, the more we see revenue growing, the more we see demand growing, the more we want to invest to be able to serve that on this 18 to 24 month lead time. So the fraction of revenue that goes into Trainium will – that will come down, but the total amount of capital that goes into Trainium, I expect that to continue to increase, but I expect what Andy said to really hold true. Now, as the industry matures, like, you know, when Andy says 30 to 40% less expensive Trainium, that's a big deal. Everyone's very focused on how we're going to bring costs of delivering these services down. You know, people joke in our industry about intelligence too cheap to meter, but we want to get as close to that dream as possible. And getting very capital efficient on delivering these services at huge scale is how we see this working.
SORKIN: And back to this exclusivity idea in terms of this Statefulness effort. How long will this be with Amazon? Is it possible you're going to want to do this – I imagine other people are going to want to do this with you. I mean, this is sort of the next frontier of what agentic AI looks like, Sam.
ALTMAN: Yeah. We – Andy can talk about timelines and stuff, if he'd like to, but we want to have a long, deep relationship here. We've done this in the past. You know, we did and still have a long, stateless API exclusivity with Microsoft. And there's a nice thing about aligning with a partner on a piece of technology. And this is a place where I think an approach where we're really going to go together and do this together will be great for both companies
SORKIN: Andy, you know, one of the things I imagine people are going to be asking this morning is, you've had a long relationship with Anthropic and Claude, and you had an early investment with them. How does this change that dynamic if it does at all?
JASSY: We still have a very strong relationship with Anthropic. Remember, we have lots of customers who run cloud models in bedrock. Anthropic has been early and very substantial customers and users of Trainium. They're training their next version of Claude, as we speak, on top of Trainium. And we have a significant investment in Anthropic as well. And they've always had multiple partners, and we do too, and so that relationship will stay strong, and we're really excited about the partnership we're building over a long period of time with OpenAI.
SORKIN: Hey Sam, while I have you here, just because I know everybody's fascinated by it and it's sort of the story in the news this morning, beyond this big news that you've just announced, which is this debate between the Pentagon and anthropic about how to use these large language models and whether they can be used improperly and what kind of rules should be placed around them. What do you think?
ALTMAN: Look first of all, the government, the Pentagon needs AI models. They need AI partners. This is like clear and, you know, I think Anthropic and others have said they understand that as well. I don't personally think the Pentagon should be threatening DPA against these companies, but I also think that companies that choose to work with the Pentagon, as long as it is going to comply with legal protections and the sort of the few red lines that the field we have – I think we share with Anthropic and that other companies also independently agree with. I think it is important to do that. I've been – for all the differences I have with Anthropic, I mostly trust them as a company, and I think they really do care about safety, and I've been happy that they've been supporting our warfighters. I'm not sure where this is going to go.
SORKIN: In terms of OpenAI, and where you go, $730 billion is a huge valuation, and $110 billion provides you what runway in terms of when you think about the timing of what that capital gives you, and then to the extent that you may think you need more capital in the future.
ALTMAN: I don't know the exact date, but it's a very long runway. I expect we will need more capital. I don't know what the form will be. It may not – you know, there's, like, a lot of ways we could continue to finance growth, but this is like, a long runway from now.
SORKIN: Related to that – and maybe Andy can speak to this, the other big story of the morning, which relates very much to AI, and I think we're all trying to understand it, you know, Andy, you and I have talked about jobs for a long time in AI. Jack Dorsey just announced a plan that effectively cuts 40% of his staff because he believes that AI is going to allow him to be just as productive, if not more so. And I wonder what your reaction is to that, and what you think it portends for the rest of us, if you will.
JASSY: Well, I haven't really digested that news very much, and I think every company is going to make their own decisions on these things. And you know, my view of it is what we've talked about in the past, Andrew, which is just, you know, I think that this is the most transformational technology shift that we've seen in our lifetime. I think it's going to impact the way we all do our work. And I do believe that a lot of the jobs that we've thrown human beings at the last 20 or 30 years, you won't need as many human beings doing those same jobs. But I also think there will be other jobs created. And that has always happened in every technology shift. And you know, I mean, 15 years ago, there was no such thing as a cloud solutions architect, and today there are tens of thousands, maybe 100,000 plus, of these types of jobs. So we will have lots of new jobs and, you know, and there'll be some sort of transition, and we'll all work through it together.
SORKIN: Sam, before we let you go, you know, one of the things that people have always talked about is AGI. You and I have talked about AGI for a long time. One of the deals that you, or terms you had in one of your earlier deals with Microsoft, had to do with, you know, what happens once you get to AGI. So I was going to ask you whether any of those terms exist, either in this fundraising round, and sort of, how far you think we are to that now?
ALTMAN: I don't have – I can't tell you the exact date that AGI arrives. And also, you know, it's gotten to the point where people define it in different ways. The field is clearly making rapid progress. So I don't want to, like, give a timeline, but if you look at what the current models are capable of, you know, you hear people say, man, it feels like something I didn't expect to see this fast or I didn't expect to see this in my lifetime. The – I expect continued, steep progress from here. I think we are all going to be surprised about the amount of AI progress the field sees this year. Probably every year after that. So I don't have a timeline to give you, but it does feel like we are making maybe faster progress than even I expected.
SORKIN: Does that change any of the deals, though? Is there – are there these AGI term limits, if you will, in any of these transactions you're doing?
ALTMAN: The Microsoft one has one, as you mentioned, but –
SORKIN: But going forward –
ALTMAN: Even – AGI will have, like, a very strong, long relationship with Microsoft. We're not – given that we think it's like a near term thing now, it's – we're not doing new deals that, like, you know, stop when AGI gets reached.
SORKIN: Fair enough. Sam and Andy, congratulations on the deal. It is a big one, a big tectonic shift in the AI business, and we appreciate you joining us first after the deal was just announced.
ALTMAN: We are delighted to get to work together. Thank you.
JASSY: Looking forward to it. Thanks for having me.
