A company that most consumers have never heard of just pulled off the biggest tech IPO in seven years — and the number it landed on should make Nvidia’s board very uncomfortable. Cerebras Systems opened at $350 a share on Wednesday, surged 68% from its $185 IPO price, and closed at $311.07, giving the wafer-scale AI chipmaker a market cap of roughly $95 billion. The offering itself raised $5.55 billion, making it the largest U.S. tech debut since Uber’s stumble onto the public markets in 2019. Except this time, the stock actually went up.
The Numbers That Made Wall Street Lose Its Mind
Demand for Cerebras shares was more than 20 times oversubscribed. That’s not a typo. For every share available, twenty investors were fighting over it. The company had already bumped up its IPO price range during the roadshow — from an initial target of $140–$165 to $170–$185 — and still priced at the top. When a deal is that oversubscribed, the first-day pop isn’t a surprise. The size of it is.
What’s backing that demand isn’t hype — it’s a financial transformation that reads like a textbook turnaround. Cerebras posted $510 million in revenue in 2025, up 76% year-over-year. More importantly, it swung from a net loss of nearly $500 million to a net income of $237.8 million. In the AI chip space, where most startups are still hemorrhaging cash and begging for their next funding round, Cerebras walked onto the Nasdaq already profitable. That distinction matters more than any valuation multiple.
Why This IPO Is Really About Nvidia’s Grip Loosening
Here’s the part that gets buried in the ticker-tape coverage: Cerebras doesn’t compete with Nvidia the way AMD or Intel do. It doesn’t make GPUs. It makes wafer-scale engines — entire silicon wafers turned into single, massive chips designed specifically for training and running large language models. The architecture is fundamentally different, and for certain AI workloads, it’s meaningfully faster.
The market just valued that architectural bet at $95 billion. For context, AMD’s entire market cap hovers around $220 billion. Intel’s is under $130 billion. A company that shipped its first commercial product a few years ago is now worth more than half of Intel. That’s not irrational exuberance — that’s the market saying it believes the AI compute stack is about to fragment, and Nvidia’s 80%+ share of training hardware isn’t permanent.
Cerebras’s secret weapon is its anchor deal with OpenAI, reportedly worth north of $20 billion over multiple years. When the world’s most prominent AI lab is willing to write that kind of check to a company that isn’t Nvidia, it sends a signal that reverberates through every procurement decision in the industry. Microsoft, Google, and Amazon have all been quietly exploring alternatives to Nvidia’s H100 and B200 chips. Cerebras just gave them proof that alternatives can scale — and go public at a near-$100 billion valuation while doing it.
The IPO Window Is Now Wide Open — And That’s the Second-Order Effect
The tech IPO market has been functionally dead since 2022. Instacart’s underwhelming debut, Arm’s slow fade after its initial pop, and a string of postponed offerings made bankers skittish and founders patient. Cerebras just kicked that door off its hinges.
Within hours of Cerebras pricing, reports emerged that SpaceX holders signed off on a 5-for-1 stock split ahead of its own expected IPO. CoreWeave, the GPU cloud company that went public earlier this year, saw its stock tick up in sympathy. The message to every AI-adjacent company sitting on an S-1 draft: the window is open, the appetite is real, and the premium for “AI” in your pitch deck has never been higher.
But there’s a catch. Cerebras earned its valuation with real revenue, real profits, and a real customer in OpenAI. The companies that try to follow it through the IPO window with nothing but a pitch deck and a prayer are going to get punished. Wall Street is euphoric about AI, but it’s a discriminating euphoria — it wants proof, not promises. Cerebras delivered proof. The question is whether the next ten companies in the pipeline can say the same.
The Risk Nobody’s Talking About
There’s one number in Cerebras’s financials that should give investors pause: customer concentration. The OpenAI deal is enormous, but enormous cuts both ways. If OpenAI shifts strategy, renegotiates terms, or builds its own chips (as persistent rumors suggest), Cerebras’s revenue base narrows dramatically. The company has other customers, but none at the scale that justifies a $95 billion market cap on its own.
Then there’s the competitive moat question. Nvidia isn’t standing still — its Blackwell architecture is already shipping, and its CUDA software ecosystem remains the single stickiest lock-in in all of enterprise tech. Cerebras’s hardware advantage means nothing if developers can’t easily port their workloads to it. The company has invested heavily in software compatibility, but switching costs in AI infrastructure are measured in engineering months and retraining cycles, not dollars.
The Verdict
Cerebras’s IPO isn’t just a financial event — it’s a structural inflection point for the AI chip industry. For the first time, the market has put a massive, liquid, public-market valuation on the thesis that Nvidia’s dominance has an expiration date. Whether Cerebras itself can sustain a $95 billion valuation is debatable. Whether the world it represents — one where AI compute is multi-vendor, architecturally diverse, and no longer bottlenecked through a single company in Santa Clara — is coming? That’s not debatable at all.
The biggest tech IPO since Uber didn’t just ring the opening bell. It rang the alarm.