Davos 2026: OpenAI CFO Sarah Friar 1/21/26

0:03What we're finding is consumers are going direct to ChatGPT, right? 800 million people. When I was here a year ago, we weren't even quite at 300. So we have over 3X'd in just a year. OpenAI Chief Financial Officer Sarah Fryer on set with our team at the World Economic Forum in Davos, Switzerland. It's a pretty exciting time for AI. You can't walk around the conference floor without talking about it. The OpenAI exec on staying ahead of the AI curve and on the company's newest revenue streams. We have over 5 million developers developing on the platform. commerce and into ads is yet another way to think about the 95 % of users who today access for free.

0:45Plus, what moral compass guides this powerful technology as it innovates? OpenAI's North Star is AGI for the benefit of humanity. We're a research lab. That's what we goal on. And it's not for the benefit of humanity who can pay. I'm CNBC producer Cameron Costa. Squawk Pod reports from Davos 2026. OpenAI CFO Sarah Fryer begins right now.

1:18It's another day in Davos. Decision makers in every industry and many governments are trudging through the snow and the wind in the Alps for an annual meeting that we, as a news organization, watch very closely. The news is revolving around what's happening here at the World Economic Forum today. Joe Kernan, Becky Quick, and Andrew Ross-Sorkin are there too, interviewing and meeting with CEOs and world leaders, making that news happen. Sometimes that news happens right on our set, outside in the crisp alpine air. In this next interview, Joe, Becky, and Andrew sat down with OpenAI CFO Sarah Fryer.

2:04In the meantime, we are talking all things AI here in Davos and Sarah Fryer is OpenAI's CFO. Thank you for being here. How do you now start to see or do you see sort of a distinction between your model, your models I should say, what Google is now doing, what Anthropics is doing, what XAI is doing? I mean it does seem like there's starting to be more differentiation in sort of where you think ultimately your market is. Yeah, so I think, I mean, it's a pretty exciting time for AI. You can't walk around the conference floor without talking about it. A year ago, we were starting to talk about agents and about things that could come into your life as a consumer and do tasks for you, be with you in the workplace, help you really do your job.

2:48And now we are here talking this week about the agentic enterprise. So really going deep on transformation. We think that consumer and enterprise are symbiotic. So we need to be in both. And in particular though, I think what Demas, I talked to him earlier today, Dario was just with, we all agree there is a massive capability overhang that's occurring. There is AI in people's hands today, they're using it quite simplistically, but the power users are using it seven times more than just the average user. They're coding, they're doing deep research tasks, they're in research labs causing breakthroughs, and a lot of people are still just doing kind of search.

3:26So I think that all of us are focused on that. How do we close that capability overhand? Big question about this. This is an iPhone and Apple just made a deal with Google to use Gemini in its phone. You had a relationship and still do with Apple. You were the first though. What was your reaction to learning that Gemini is now going to be effectively, for better or worse, or for lack of a better word really, the new Siri? So actually we are still baked into the Siri piece, but I think importantly what we're finding is consumers are going direct to ChatGPT, right? 800 million people. When I was here a year ago, we weren't even quite at 300, so we have over 3X'd in just a year.

4:08It's kind of wowing. If you look at the actual use of Genii, so it's not just about users, but it's how deep are they going. They are absolutely doing that inside ChatGPT. We've become the noun and the verb. So in a way, we view the iPhone, and I just saw Tim, as amazing distribution partners. But we think we have to be broad on that distribution spectrum because users want to be able to use their technology on whatever device they have. But you don't think Gemini now has an advantage by having the surface area not just of Android but of iPhone in that way? No, because I think people are coming direct to ChatGPT as the brand they trust from an AI perspective.

4:45And we also have many other partnerships and distribution channels that we're working on and have announced. But again, going back to how do we shift consumers out of just a chat GPT call and response search, like what's the weather in Davos today, into something deeper like a task worker. So we announced healthcare is a great example where you can go deep. And now suddenly you have it working for you to help you maybe look at a biopsy and give you a second opinion. And we see in health deserts that people are using ChatGPT to help them get some guidance. And even 66 % of U.S. physicians say they're using ChatGPT daily in their work.

5:27I find that amazing and so hopeful for what's available to us. I do too. I think healthcare is probably the thing where I have the most hope for what we see from AI. But when Dario was here, he was talking about the advantages that they have at Anthropic. One of them is that he is 80-20 enterprise versus consumer. And he thinks that enterprise is a more reliable source of money, and it's also a place where you see higher margins. What are you doing besides this health care initiative to try and make sure that you lock your consumers in and that they're more willing to not only use it more, but also pay more for it?

6:00Yeah, so maybe just talking about our business, at the end of last year, sitting right here, I would have said our business 50-50 consumer enterprise. Or sorry, 70-30 consumer enterprise. Today, that number 60-40, I think we'll end the year closer to 50-50. So again, while the consumer business has been growing at an extraordinary rate, literally the largest growing consumer platform the world has ever seen, our enterprise business is also the fastest and largest in AI the world has ever seen. One million businesses using the platform today. how do we think about building consumer trust and loyalty like I hate to talk about lock-in because it really should be about value I think on the consumer side you have value you lock them in exactly on the enterprise side how do we drive that kind of moat I mean number one is enterprises are starting to move from maybe a chat GPT wall-to-wall deployment into big transformations I was just with the CEO of a bank who I was there to pitch him and ironically he had just sat with Carlos Torres, who is chairman of BBVA, and my work was done.

7:06Because Carlos said, hey, BBVA, we have seen transformation. Like OpenAI helped us go from 10 ,000 deployed seats to over 120 ,000. They have helped us deploy into our call centers, into our credit screening. It is truly transforming the bank. We're in 25 countries. We can do all that in a multilingual way. And so that is where I think when you're that deep in a business, they are loyal because they're getting so much value. On the consumer side, you were doing something that Sam for a very long time said he did not want to do. It was almost going to be a last resort situation. Advertising. Implementing advertising.

7:45How are you thinking about that? And how do you do it in a way where the customer trusts the answers that are coming back from ChatGPT when you're asking it, plus effectively what the advertisers are paying for? Yeah. So let me talk about the North Stars of advertising as we begin to do some testing. But then I want to talk about the optionality that we're trying to build, broadly speaking. On the trust side, number one is what you said. North Star is the model always gives you the best answer, not the paid for answer. That has to remain goal one. Number two is we don't share your conversations with advertisers.

8:18We don't sell your data to advertisers. But I think number three is offering people a non ad layer as well or on ramp. But why are we doing it? It comes back to access, right? Open AI is North Star is AGI for the benefit of humanity. We're a research lab. That's what we go on and it's not for the benefit of humanity who can pay. And so in order to drive access we need to drive more multi-dimensionality on the infrastructure. So sometimes it might be for a frontier model, right? Every six months we see a leap in that intelligence. Or it might be on the inference side, so we drive down costs and are able to supply to users at a lower cost.

8:55At the product level, multiple products. And then at the business model level, which is where you get to ads, we started with subscriptions and now we are multi-subscription. We have a SaaS-based model for enterprises as well as API. Remember, we have over 5 million developers developing on the platform. Commerce and into ads is yet another way to think about the 95 % of users who today access for free. And then finally, I think there's an interesting value exchange model in our future where we align ourselves on the same side of the table as, say, a drug developer. Let's say it's Amgen, it's Eli Lilly, it might be a group like Thermo Fisher, and we say we're going to work with you to create this next breakthrough.

9:35Can we take a license off of the outcome? But can I just ask you, I appreciate the North Star as a consumer. I don't want you giving me results based on somebody paying for them. But if I'm an appetizer, why am I paying you if it's not to get a better result? So number one, what we see happen in ChatGPT is a lot of people are using it for commerce already. They are probing. Again, if you look at search data, a Google search is just like a one line. I'm looking for a stroller from I'm expecting a baby. What stroller do I get? In chat GPT, you see a whole conversation. You know, I live in San Francisco.

10:10There's lots of hills. What's the safest one that's not going to roll down a hill? I'm having twins. I need a whole different setup. So first of all, the conversation is really meaty and gets down into consideration. At that point, we may enable an advertiser, a consumer, very clearly to see you're now going to go talk to an advertiser, right? I need a place to stay in Davos this weekend. Maybe Airbnb is the perfect place to go have that conversation. But it's going to be clearly labeled. Clearly labeled that you've moved into a new zone and that you're now talking to an advertiser. But that's why we see even in ChatGPT, right, when we open up the platform to bring in a Booking.com or an Expedia or an Airbnb, so much excitement because users are saying, OK, I got to that moment of consideration, but now help me get to the purchase.

10:55Right. Before I let you go, obviously, there's a lot of speculation about when you're going to go public, what's going to happen. But before that happens, Elon Musk is now continuing his lawsuit and it appears that a trial could happen here now. So tell us just about what you think could happen over the next couple of months, if not the year. I know you're fighting, obviously, to remain under this new structure, but there's a whole lot of things that have to happen and go right or wrong before you can get to an IPO at this point. I mean, first of all, we're an incredibly strong business today.

11:27Again, in just a very short two and a half year period, we've gone from zero to a company that is now a PPC, a public benefit corp. So when it comes to things like fundraising, got a heck of a lot easier. Secondly, we have an incredibly strong balance sheet. We exited this year or 2025 with over$40 billion on the balance sheet. We executed the largest private fundraise ever last year,$41 billion raised. And then as we look forward, for us it's eyes on the prize of how do we add value to the consumer, how do we add value to enterprises, and how do we close this capability gap. You're going to hear us talk about this over and over again.

12:02But does the lawsuit itself, I mean if you were to lose the lawsuit, would change the entire business model? We just try not to get distracted. Elon has launched many lawsuits at us, even since I joined the company. And in the end we think that they're spurious, that it's a distraction tactic. And for us, we've done an incredible job, I think, of continuing to just execute through that noise. Because in the end, we need to show up for our customers. Okay. Sarah Fryer, it's great to see you. Thank you. Nice to see you all.

12:39Thank you for listening to this special Squawk Pod reports from Davos, Switzerland at the World Economic Forum. As we do every year we go to Davos, we are bringing you the best interviews from our entire week there right to your SquawkPod feed. Make sure you follow us wherever you're listening now. And then those episodes, like interviews with President Donald Trump and Treasury Secretary Scott Besant, they'll pop up automatically every time we post. Squawk Box is hosted by Joe Kernan, Becky Quick, and Andrew Ross Sorkin in Davos or, more typically, in Times Square in New York. Squawk Pod is produced by me, Cameron Costa, and Zach Valisi.

13:20Our Katie Kramer is producing on the ground in Switzerland this entire week. Thanks to Julie Trass, our editor for the Davos episodes. Catch you next time. Have a great day.