AI Chipmaker Cerebras’ CEO Is Now Worth $3.4 Billion After Long-Awaited IPO

1 month ago 4

Cerebras CEO Andrew Feldman

Sportsfile via Getty Images

In March 2025, as Cerebras’ first attempt at going public sat in limbo, CEO Andrew Feldman stood in front of a crowded room at the Glasshouse, a trendy party venue in downtown San Jose, California, raising a glass of champagne. He was in the proverbial belly of the beast, holding a swanky event on the sidelines of Nvidia’s GTC conference, schmoozing potential customers in town for the chip giant’s massive annual summit.

“I’d like to say cheers, and enjoy the party,” he said. As he finished his toast, a curtain opened to reveal a dance floor, disco ball and smoke machine, while Justin Timberlake’s “Rock Your Body” blasted on the speakers.

It was a posh gathering, where guests nibbled on shrimp cocktails and avocado mini tacos. The elaborate evening, including the surprise dance floor reveal, seemed illustrative of something Feldman told Forbes earlier that day when asked about competition with Nvidia: “The question is not, are you competing?” he said, though his answer was an emphatic yes. “The question is, do you have to take share from them, or is the market growing so fast that everybody can get fed?”

In other words, there was plenty of room for he and Cerebras to grab their share — even if, like the hidden dance floor, it wasn’t apparent at first.

More than a year later, as the biggest AI companies in the world commit trillions of dollars to chips and data centers, his sentiment is only more true. On Thursday, Cerebras finally went public. The company raised a whopping $5.55 billion at a $185 per share offering price. That was already enough to make Feldman a billionaire, thanks to his roughly 5% stake. Then shares opened trading at $350, pushing the value of his shares and options to $3.4 billion. His cofounder Sean Lie, Cerebras’ CTO, joins him in the ranks, with stock and options worth around $1.9 billion.

The company makes chips that it claims are faster and cheaper than Nvidia’s industry-standard GPUs when it comes to inference — the computing processing power it takes to run AI models, instead of training them. Cerebras’ specialty is its wafer chip, a large slab of hardware about the size of a dinner plate, roughly 58 times bigger than an Nvidia GPU. Cerebras argues that its giant chips work better for memory bandwidth and chip-to-chip communication than other silicon on the market.

The road to Cerebras’ public market debut has been long and windy. The company first filed its S-1 in September 2024, revealing that G42, a United Arab Emirates tech conglomerate, accounted for 87% of revenue for the first half of that year. The process hit a snag after G42 investment drew the scrutiny of CFIUS, the government committee that reviews foreign investment in the U.S., which then opened an investigation over national security concerns. The company reportedly canceled its roadshow and postponed the IPO. After months of uncertainty, Cerebras eventually passed CFIUS clearance in March 2025.

But instead of immediately resuming its path to an IPO, the company instead turned to private markets to build its war chest. Last September, it raised $1.1 billion at an $8.1 billion valuation from backers including Fidelity Management, Valor Equity Partners and Benchmark. Then in February, it hauled in another $1 billion, led by Tiger Global, ballooning its valuation to $23 billion. Then last month, Cerebras filed a new S-1, reporting revenue of $510 million in 2025, up nearly 76% from the year before.

The son of biology and psychology professors at Stanford, Feldman, 56, was literally raised on the world famous campus. “When you grow up on a university, you're pretty clear on what brilliance looks like,” he says. “The guy two doors down the road had two Nobel Prizes, and all we knew was that his wife gave out full-sized candy bars on Halloween.” He went on to attend Stanford himself, first as an undergrad in 1987 studying economics and political science, then as an MBA. In 2007, he cofounded a server company called SeaMicro, which sold to AMD in 2012 for $355 million. After two years at the chip giant, he left and cofounded Cerebras in 2016.

For Cerebras, the hope is that the seemingly insatiable thirst for compute from big labs will propel its business forward. Google has earmarked up to $190 billion for compute this year, more than double the $90 billion it spent in 2025. Meta said it plans to spend up to $145 billion. Project Stargate, an effort by OpenAI, SoftBank, Oracle and others to build a planned $500 billion worth of AI infrastructure, has been a marquee tech initiative to kick off President Trump’s second term, though some progress on the effort has reportedly stalled. All told, big tech could pour an estimated $500 billion into AI data centers and chips this year alone, according to a report by Goldman Sachs.

Cerebras is just one of several next-gen chip companies on the market, but it’s the first of its ilk to go public. Groq, started by former engineers who worked on Google’s TPU chips, sold to Nvidia last Christmas for $20 billion, becoming a key part of the chip giant’s inference strategy. SambaNova, another competitor, was reported to be in acquisition talks with Intel in December, before changing course and raising investment from the chipmaker and others in February at a $5.1 billion valuation.

Meanwhile, Cerebras has bolstered its market position further ahead of its IPO with a handful of high-profile deals. In January, OpenAI agreed to buy up to 750 megawatts of compute from Cerebras over three years in a deal valued at more than $20 billion. Cerebras also partnered with Amazon in March to bring its chips to the ecommerce giant’s AWS data centers.

Now as a public company, Cerebras is starting a new chapter. Last year, as the company was mired in its CFIUS drama, Feldman discussed his biggest concern as the company continued to grow. “I think the incentives in large companies are not to do pioneering innovation. The cost of failure, they believe, is very high,” he said. “We want massive return and massive and big risk. That's what we're looking for.”