The Takeaway
- OpenAI is committing billions of dollars years in advance to help finance data centers.
- Friar last year began reporting to the head of applications instead of to Altman.
- Altman has excluded her from some conversations related to financial plans.
Sam Altman has committed OpenAI to spend $600 billion in the next five years and privately said he wants to go public as soon as the fourth quarter despite expectations his company will burn more than $200 billion before it starts generating cash.
Behind the scenes, Sarah Friar, his chief financial officer, has voiced concerns that reflect the tensions and risks inherent in the CEO’s extraordinarily ambitious plans.
She told some colleagues earlier this year that she didn’t believe the company would be ready to go public in 2026, because of the procedural and organizational work needed and the risks from its spending commitments, according to a person who spoke to her. She said she wasn’t sure yet whether OpenAI would need to pour so much money into obtaining AI servers in the coming years or whether its revenue growth, which has been slowing, would support the commitments, said the person who spoke to her.
It isn’t clear whether the company’s announcement this week that it had received investment commitments totaling $122 billion has allayed some of her concerns. OpenAI is due to receive the capital in stages, and it mostly will come from Amazon and Nvidia, which supply the cloud servers and chips OpenAI uses. The deal is part of a series of circular financial arrangements involving the biggest AI-related companies.
Outside OpenAI, Altman, 40, and Friar, 53, have generally presented a united front. That was the case at a dinner for investors that they hosted at Altman’s San Francisco home earlier this year, said one attendee.
At times, though, their divergent emphases have fueled the perception of strains in their relationship among some people who have worked closely with them since Friar joined OpenAI in June 2024.
Those people said Altman has excluded her from some conversations related to the company’s financial plans. For instance, in recent months he left Friar out of a conversation about server spending with leaders at one of OpenAI’s top investors, one of these people said. Her absence was noticeable and awkward, given that a previous conversation on the same topic included her, according to an attendee.
A different person who attended a senior-level meeting at OpenAI with Altman earlier this year said it was unusual that Friar was not invited, as it involved a discussion of major financial decisions.
In an unusual move for a large company, where CFOs almost always answer directly to the CEO, Friar stopped reporting directly to Altman in August last year and instead began reporting to Fidji Simo, who had joined as head of OpenAI’s applications business. Simo last week told staff she would take a short medical leave.
Upgrade to ask Deep Research to…
What new assets are bankers considering securitizing to fund AI data centers?
What’s the latest on SpaceX’s IPO?
How is AWS using AI to automate roles previously handled by staff?
How do OpenAI and Anthropic differ in how they report revenue?
A spokesperson for Altman and Friar sent a joint statement from the executives: “We are fully aligned that durable access to compute is at the core of OpenAI’s strategy and a key differentiator as we scale. We have both been directly involved in every consequential compute decision over the past year plus. The $122 billion round locks in the capacity to scale compute aggressively and positions us to become the core infrastructure layer for AI, translating that advantage into sustained leadership across research and products, and making it possible for people around the world and businesses, big and small, to just build things.”
‘A Hard Job’
Altman and Friar have starkly different personalities and backgrounds. He has long positioned himself as a world-changing visionary in the mold of Steve Jobs and Elon Musk.
She is a former Goldman Sachs equity analyst who later worked in finance at Salesforce, helped Square (now Block) go public and spent around six years as CEO of Nextdoor, a local social network, before being replaced as the business struggled. She was then brought in to OpenAI to help it raise funds and reassure investors about what Altman has described as the most capital-intensive company in history.
Friar “has a hard job,” said someone who works closely with her and Altman. “She is working for a founder with big ambitions who wants to push the envelope as hard as he can on spend.”
Other Silicon Valley startups have experienced friction between visionary, growth-obsessed founders and risk-averse finance executives brought in to prepare a company for public markets.
One of the most prominent clashes occurred at Airbnb between CEO Brian Chesky and then-Chief Financial Officer Laurence Tosi over the company’s business strategy and organizational structure. Tosi ultimately departed the company in early 2018, creating a significant leadership vacuum just as Airbnb was preparing for an anticipated IPO, contributing to a broader exodus of executives who were frustrated by the delayed liquidity event. (Chesky successfully took the company public two years later and also became an informal adviser to Altman.)
OpenAI has had fissures between top executives in the past. In late 2023, two senior executives helped to oust Altman before he negotiated a return. Years before then, several senior leaders fell out with Altman and left to start Anthropic.
Even with OpenAI’s recent fundraising, Altman and Friar will need to continue to convince investors to hand over tens or hundreds of billions of dollars in capital, and pressure is rising to meet revenue and user goals to support its rising cash burn. Anthropic has become a major threat and already surpassed OpenAI in selling AI models to businesses and app developers, while Google’s Gemini has been chipping away at ChatGPT’s dominance of the consumer chatbot market.
While OpenAI recently increased its revenue outlook for the next five years by 27%, it privately warned investors in February that it will burn more than twice as much cash through 2030 as previously predicted.
In another sign of its financial pressures, OpenAI told investors that its gross profit margins last year were lower than projected due to the company having to buy more expensive compute at the last minute in response to higher than expected demand for its chatbots and models, according to a person with knowledge of the presentation. (Anthropic has experienced similar problems.)
Talking to Goldman
Despite the resistance Friar has expressed to fast-tracking the IPO, OpenAI has begun to lay the groundwork to go public by tapping law firms Cooley and Wachtell Lipton Rosen & Katz and having informal conversations with IPO bankers at Goldman Sachs and Morgan Stanley. Amazon and Nvidia could also sway the timing of a public offering, given their now-sizable interest in the firm. (Altman has privately said he wants OpenAI’s public offering to happen before Anthropic, which is discussing an IPO in the fourth quarter of this year.)
If it moves forward, the IPO would be one of the largest in history, and a monumental undertaking for Friar’s finance team and many other groups.
OpenAI in some cases is committing billions of dollars years in advance to help finance the construction of specially designed data centers for developing new AI models.
Publicly, Friar has played down the notion that Altman is steering her into uncomfortable territory.
“I do get a lot of prompts and pushes from Sam…They definitely sometimes can make me feel like, ‘Oh, okay, I have a lot to do now,’” she said in an interview with the Wall Street Journal in November last year. “But they more inspire me than create anxiety because I’m here at this table because I genuinely think we’re doing something that’s never been done in the world before.”
At the time, she also said that an IPO “is not on the cards right now” because OpenAI was still trying to “get the company into a state of like constantly stepping up to the scale that we’re at.”
Sarah Friar at the Allen & Co. conference in 2023, via Getty.
Several OpenAI employees said they think the company’s aggressive server spending plans will help it stay ahead of rivals.
After The Information reached out for comment for this article, Friar on Saturday posted photos of herself and the finance team on LinkedIn with a caption congratulating them on completing the firm’s $122 billion fundraising.
“Cheers to Sam [Altman] and [President] Greg Brockman for seeing it early and acting decisively,” the post said. “The foresight sets us up for long term success.”
Cloud Computing Risks
Altman has said that getting more computing power will help the company produce better technology and ensure it has enough capacity to serve its growing user base. Executives have long complained that OpenAI is constrained by compute and has had to slow the rollout of new products and features as a result.
After Friar became CFO, Altman tried to reduce OpenAI’s reliance on renting chips from cloud providers, announcing a joint venture with Oracle and SoftBank to develop $500 billion worth of data center capacity in the U.S. OpenAI ultimately shelved its plan to own and operate facilities after lenders balked at backing multibillion-dollar projects tied directly to a company with an unproven business model that was burning billions in cash annually.
Instead, at Altman’s direction, OpenAI struck a series of deals to spend more than $600 billion renting servers from Microsoft, Oracle, Amazon and other cloud providers over the next five years.
OpenAI is “utilizing our [cloud] partners because it’s a way to keep a lighter weight on the balance sheet,” Friar said in an interview earlier this year.
Still, OpenAI in some cases is committing billions of dollars years in advance to help finance the construction of specially designed data centers for developing new AI models.
That’s different from traditional cloud computing contracts that require customers to pay only for the server capacity they use, or to guarantee a certain spending amount over the course of a year. Typical cloud customers would rely on firms such as Amazon to build enough data centers rather than need to spur those firms into action.
Friar has said AI data center development creates a natural pressure because it takes years to construct the facilities OpenAI needs, and the company has to essentially pre-order the capacity.
“Now, there is definitely a timing mismatch, because I have to make decisions today about making sure we have compute in not even [20]26 or ‘27, but ‘28, ‘29 and ‘30,” Friar said on an OpenAI podcast earlier this year. “If I don’t put in orders today and don’t give the signal to create data centers, it won’t be there, right?”
In one case, OpenAI agreed to help fund Oracle’s data center development, The Information has reported, by sharing some of the economic risk. If expenses go over budget, the companies agreed to share in paying the overage costs, for example. It’s unusual for a cloud customer to agree to such a deal term, according to people involved in the deal.
Amodei’s Swipe
OpenAI’s $600 billion in commitments aren’t set in stone, however. OpenAI could slow down the construction of some facilities, and therefore its payments, or back out of some deals, two people with direct knowledge of OpenAI’s compute contracts said.
Last September, Friar recruited Mike Liberatore, who was CFO of Musk’s xAI, now part of SpaceX, to work with her and Brockman on server-related finance deals, according to two people with knowledge of his role.
Friar’s earlier comments to colleagues about her concerns regarding spending are similar to statements that Altman’s archnemesis, Anthropic CEO Dario Amodei, has made.
“Even if the technology goes as fast as I suspect that it will, we don’t know exactly how fast it’s going to drive revenue…. But with the way you buy these data centers, if you’re off by a couple years, that can be ruinous,” he said in a Dwarkesh Patel podcast in February.
“If I’m just off by a year…or if the growth rate is five times a year instead of 10 times a year, then, you know, you go bankrupt,” Amodei said, adding, in an apparent swipe at OpenAI: “I kind of get the impression that…some of the other companies have not written down the spreadsheet, that they don’t really understand the risk they’re taking.”
