There’s one thing top of mind for CFOs at A.I. companies: compute power. As their firms race to dominate Silicon Valley—and trounce one another—OpenAI CFO Sarah Friar and Anthropic CFO Krishna Rao are focused on raising funds, securing chip deals and tracking demand for compute.
Compute power is a “huge competitive advantage,” especially in a landscape where “there’s not a lot of compute in 2026,” said Friar, 53, in an interview with Bloomberg TV today (May 15). Securing enough capacity to meet demand is particularly challenging for OpenAI, which has more than 900 million weekly active ChatGPT users and an enterprise sales team “run ragged” by businesses eager to adopt its models.
Before joining OpenAI, Friar was CEO of Nextdoor from 2018 to 2024 and spent six years as CFO of Square, following roles at Salesforce, Goldman Sachs and McKinsey.
Anthropic, OpenAI’s primary rival, is equally focused on securing chips. “The compute that we procure, it’s the lifeblood of our business—it is the most important thing in the company,” said Rao, 43, on an episode of the Invest Like The Best podcast published May 13. He added that he spends 30 percent to 40 percent of his time on compute-related decisions.
Like Friar, Rao is about two years into his current role. He previously served as CFO of Fanatics and Cedar, after a six-year stint at Airbnb and earlier work as a senior associate at Blackstone.
Both executives face a delicate balancing act: deciding how much compute to secure as their companies scale rapidly. These are “some of the most consequential and hardest decisions” a business can make, Rao said. Too much compute risks overspending; too little leaves customer demand unmet. “You have to really think ahead to plan this.”
To meet their needs, both companies rely on a mix of suppliers. Anthropic primarily uses Nvidia GPUs, Google TPUs and Amazon’s Tranium chips. OpenAI also works with Nvidia and Amazon, while incorporating AMD chips, developing in-house hardware with Broadcom and deploying compute from Cerebras Systems, which went public yesterday (May 14).
The high cost of A.I. infrastructure has driven massive fundraising. “It’s a capital-intensive business,” said Rao. When he joined Anthropic, the company was closing a $750 million round at a $15 billion valuation. It is now reportedly seeking a $30 billion round that could value it at $900 billion—potentially surpassing OpenAI, which was valued at $852 billion in March following a $122 billion cash injection.
OpenAI is also considering additional fundraising. While its recent twelve-figure raise “gives us a lot of optionality,” Friar said she isn’t “averse to raising more capital.”
Founded a decade ago, OpenAI has long led the A.I. race. But Anthropic, launched in 2021 by former OpenAI employees, is rapidly closing the gap. Its revenue run rate surpassed $30 billion in April, up from $9 billion at the end of 2025. OpenAI reported $24 billion in annualized revenue as of March, though the figures aren’t directly comparable, since OpenAI doesn’t include cloud partner sales.
Both companies are also moving toward public markets in what could be two of Silicon Valley’s most anticipated IPOs. Anthropic’s offering could arrive by the end of the year, while OpenAI’s may come in the second half of 2026.
Despite calling private markets “incredibly generous to us,” Friar said public markets offer greater scale over time. “We want to be able to, over the long run, raise money across the whole spectrum—and the public markets are significantly bigger.”
