OpenAI May Raise More Capital as Compute Crunch Deepens, CFO Says

🎥 May 15, 2026 📺 Bloomberg Television ⏱ 11m 👁 5027 views

OpenAI CFO Sarah Friar says the ChatGPT maker may raise more capital even after completing what she described as the largest private fundraising round ever. She speaks with Caroline Hyde on Bloomberg Television. -------- More on Bloomberg Television and Markets Like this video? Subscribe and turn on notifications so you don't miss any videos from Bloomberg Markets & Finance: https://tinyurl.com/ysu5b8a9 Visit http://www.bloomberg.com for business news & analysis, up-to-the-minute market data, features, profiles and more. Connect with Bloomberg Television on: X: / bloombergtv F...

Watch on YouTube

About Sarah Friar

Sarah Friar, CFO of OpenAI, has stated that the company raised over $120 billion in March 2026, describing AI as "the biggest era that we've seen today" and expressing a belief that it will lead to greater global productivity and affluence. Regarding a potential initial public offering, Friar has characterized an IPO as "a milestone" and "not a destination," adding that it is "just another way to fund raise" and that her role is to "create optionality." She has noted that compute remains a "very scarce resource" and that demand is outpacing supply, stating that the company's earlier investments in compute capacity were prescient. In discussions about AI's impact on finance and careers, Friar has said that AI will not take away all jobs but will change many, with some routine roles likely being replaced by agents. She has described OpenAI's finance team as lean, with around 200 people, and has highlighted the use of custom GPTs for tasks such as investor relations and audit support. Friar has emphasized curiosity, adaptability, and kindness as key skills for the future workforce, and has stated that human trust and relationships remain central to fundraising and business. She has also described ChatGPT as having "very high intent" for advertisers, comparing its potential ad platform favorably to those of Google and Meta.

Source: AI-verified profile updated from Sarah Friar's recent appearances. Browse all interviews →

Transcript (24 segments)

Interviewer0:00

We're just talking about the euphoria in the market again. Anything related to compute, anything related to investment opportunities, how much do you still need to invest in compute of your own to keep the edge?

Sarah Friar0:11

Yeah. We believe compute is a huge competitive advantage. We're going up a vertical wall of demand as we've talked about even just in the last six weeks seeing the uptick of 5.5, our newest model, 5.5 Cyber. So everyone wants to talk to us about that model in particular. And then on the Codex side, we're seeing just incredible lift, over 4,000,000 users now on Codex. That's up, I think it went from 3,000,000 to 4,000,000 in just two weeks. And so we need compute, and there's not a lot of compute in 2026. So we are looking out into future years and saying, what can we do to get that compute today? How do you afford that compute? So mostly our business model. So we have seen an incredible lift in our revenues to match just the demand that I'm talking about. On top of that, we have raised a lot of money, as you know, from an equity standpoint.

Interviewer0:59

More than 20 billion. I mean, could you need more then?

Sarah Friar1:01

It will depend somewhat on matching the demand that we're seeing, the revenue that we'll create, and hence the cash flows, and then that gap to wanting to buy compute out into future years. So I'm not averse to raising more capital, although right now that 22 billion gives us a lot of optionality, and I think derisks a lot of what we need to go do in the market at the moment.

Interviewer1:25

So dare I ask, would you raise it in the private markets? We just have reports Anthropic is looking at slightly bigger valuation than yours at $900 billion and going out and raising $30 billion at least. Would you go to the private markets, or would it be a public?

Sarah Friar1:37

The private markets have been incredibly generous to us. Much bigger to tap than I think most people ever thought possible. I mean, 22 billion is literally the largest private fundraise ever. But there is something attractive about the public markets over time too because it's not just about equity. Right? We want to be able to, over the long run, raise money across the whole spectrum, and the public markets are significantly bigger than the private market. So my job as a good CFO is always to create optionality. But, you know, first and foremost is we need to keep meeting our customers where they are, and they have a voracious demand right now for our technology.

Interviewer2:13

Talk about that demand because there had been reports that you hadn't met some of your highest aspirations, stretch goals that you'd set yourselves in terms of how many people are coming to the platform and how customers are using it. Are you meeting them now? Have you got a billion people on the platform? How are you looking at it?

Sarah Friar2:28

So still giving out our number of 900,000,000 plus weekly active users using ChatGPT. Our mission is AGI for the benefit of humanity, and our goal is always to get it out into the hands of people and get them using it. So we're delighted with how the consumer platform is going. Images too. So our most recent image model is the best in the world, and it's a ton of fun if you haven't tried it. And that's really caused a very nice uptick on the consumer side. In fact, one of the things we see is that as people come to use something like images, it becomes a really strong acquisition channel for us. And then on the enterprise side, I am, as we talked about, just back to back to back with customer meetings. I'm the CFO. So our sales team has just run ragged at the moment. Everyone wants to talk to us about what can I do to transform my business? How can I use the intelligence of the models that's available to me today? And so we just see our business going from strength to strength right now.

Interviewer3:21

From strength to strength, from a consumer perspective, you need the right partners in place. Now there was an interesting article put out by Mark Gurman today saying that maybe that partnership with Apple is slightly fraying because you're not getting the sort of treatment or at least the user case that you had expected. Maybe even you bring some legal work towards Apple within that partnership. Is that true? How is that relationship going? Would you think about bringing legal action?

Sarah Friar3:50

So I'll go back to that mission statement. So AGI for the benefit of humanity. Our goal is to find great partners that have really large distribution to get it into the hands of consumers. So when I look at, we have partners like Kakao in South Korea. PhonePe has actually been an incredible partner in India. And so we know that there is a route to market via these partners, and it can be wildly successful. And we're just gonna keep pushing on that.

Interviewer4:13

You are dodging that question. So is the relationship strong with Apple?

Sarah Friar4:19

I mean, right now, in terms of our relationship with Apple, it's like every other large distribution partner for consumers. We want to make it work, and we know that consumers want our technology. So we think that there's a way.

Interviewer4:28

You think that there's a way without or with legal action, you think?

Sarah Friar4:34

I cannot comment on legal action.

Interviewer4:37

Can you talk to us a little bit about access to compute and how that is an edge versus competition in the market? There has been thoughts that maybe it has perhaps been allowed with certain partnerships maybe because of the power of the model, or then others say maybe it's because of lack of compute. How are you thinking about rolling out your Cyber offerings and whether it's an access to compute issue or whether it's just the power of these models just being so sophisticated?

Sarah Friar5:02

So 5.5 Cyber in particular is rated very good by third party research teams. And we're really proud actually of the 5.5 Cyber model in particular. In particular, what I love about it, though, is that it's a very efficient model. So it is much cheaper for customers to be able to use it. And in keeping with this point about access, we have gone much broader in letting people have more access to it. The European Union, in fact, just made a comment earlier in the week to give us some kudos on that rollout. That doesn't happen very much. It does not happen very frequently, but it goes back to, you know, again, very mission aligned for us to make sure that we're putting our technology and democratizing it into the hands of everyone. Now in something like Cyber, we tend to be a little bit more careful because there's a lot of safety and security elements there. This is a model that really can find vulnerabilities in software code at a level that people have not seen before. And so it's also causing huge demand. I think I've been spending a bunch of time with banks and financial institutions. I have not talked to a single CEO of a bank that does not have this as their top priority currently.

Interviewer6:17

In terms of, therefore, priorities going forward with the safety of what's being deployed, but also the need for energy and for scale, what is the key bottleneck, do you think, for just getting this into as many people's hands as possible?

Sarah Friar6:32

Yeah. So there's really three big components to a model. It's the compute itself, its data, and its great researchers. The biggest compute is always the people. So great researchers, I'll always come back to. There's just not that many of them in the world. I've had that concept too. We are super proud of the team at OpenAI. I think one of the things, it's about the cost to hire them, but more than that, that they can get access to compute to do the research that really inspires them. On top of that, compute itself has clearly been a bottleneck. I don't think any of us, even Sam and Greg, who I think were incredibly prescient overall in terms of the need for compute, I don't think they could have foreseen just the sheer demand that we have right now in 2026. And so it's the energy that sits behind it. We're seeing memory out of Southeast Asia. But there's all of these choke points in the supply chain, and this is why it behooves us to get ahead of that. Strike those partnerships earlier when people are maybe not seeing just the sheer scale.

Interviewer7:32

Let's talk about partnership and about some of the people you just talked about, Greg Brockman and Sam Altman. They've had a tough week in terms of being in trial, having their relationships sort of cast into the limelight. What is this like working with Sam? There has been some questioning as to how you work alongside him. Do you work well? Is it a strong bond? Is he trustworthy? As it seemed to have been grilled by certain lawyers of Elon Musk this week.

Sarah Friar8:01

So I think Mr. Musk is very much into customers' hands, and really creating AGI for the benefit of humanity. Working with Sam is great. We get on incredibly well. I think we have a wonderful partnership, super good yin and yang. Sam pushes. He's super curious. My job is to take that curiosity and create optionality. So just as we talked about whether it's making sure we have enough compute, making sure we have the funds to do it, making sure we build the business strongly. And I spend an inordinate amount of time with customers as does Sam, and we get to compare a lot of notes on that front, as well as sometimes our notes on parenting. So it goes the whole gamut, frankly.

Interviewer8:43

How do you, when you think about the competitive landscape, it's been an extraordinary week with Anthropic turning for compute from Colossus, which is Elon Musk's data center, which serves its own Grok. How do you think of an AI stacks up at the moment versus the rest?

Sarah Friar8:58

So we're staying the course. We play our own game. I think the bigger question to ask is where are we right now overall in this industry? And I truly think we're just getting started. Right? It is literally just beginning. Right? When you talk to a customer, they're asking, what do I do? What's the first step I take? I was just with BBVA, a great example of a bank that's already deployed 130,000 seats out into all of their workforce, and yet they wanna go deeper and do more transformation. I was back in New York seeing a bunch of other banks, folks like Morgan Stanley. Morgan was one of our earliest customers on the wealth management side. And so I think that's the better question to be asking rather than is it just a, you know, is it a zero sum game? I think it's an incredibly expensive market at the moment. I think we're just learning what to do with this incredible utility called intelligence. And I couldn't be more excited about what it's gonna look like a month from now, six months, a year from now.

Interviewer9:58

You are an optimist, and, of course, you must be and you should be. And you carry a lot of responsibility with the power and the strength that OpenAI has. How do you feel about when people turn to you and say, am I going to have a job in the future? What is my life going to look like in terms of what talents, what skills I need?

Sarah Friar10:15

Yeah, so I absolutely see every trend, every technology that's come along has transformed how we work. Right? The job I have today is very different. I did this job five years ago, ten years ago. And so with that, of course, people's jobs are changing. To that end, I think we need to make sure that we're investing in the schools. So we have an incredible offering with our EDU SKU. The government of Estonia has pushed that SKU up, it's wall to wall in all of their high schools, for example. So it starts early in terms of education. I think we need to think about reskilling and other jobs platforms. We invested in that back last September. But beyond that, I think we also need to bring this optimism into the workforce and show people the upside of maybe they can focus back on the areas that bring them joy. We talked about medicine earlier and a practitioner like a doctor or nurse, they don't love their paperwork. ChatGPT can do that for them. Please. It should do it for them. And maybe we get people back into the areas that brought them a lot of joy when they first qualified into that job. But, you know, we are on a journey. I think we don't even know the jobs that are going to be created. Our chief economist just did a wonderful paper on this. And if you just look at even just new job creation in the last five years, I think five years ago, over 40% of the jobs that exist in the world today didn't exist five years ago.