Cerebras Systems files for IPO, aiming to boost AI chip market

Cerebras’ renewed filing comes on the heels of a dramatic financing surge that has vaulted the company into the upper echelons of AI‑hardware valuations. After a $1.1 billion Series G round in 2023 and a $1 billion Series H in February 2024—both at a $23 billion post‑money valuation, according to the Wall Street Journal—Cerebras now claims $2.1 billion in total capital raised. The fresh IPO prospectus, which a company spokesperson says targets a mid‑May launch, does not disclose the size of the offering, but the sheer scale of the backing suggests the firm is positioning itself to compete head‑to‑head with entrenched players such as Nvidia and AMD. The filing also marks a reversal of the 2024 withdrawal that resulted from a federal review of an Abu Dhabi‑based G42 investment, a setback that the startup appears to have fully recovered from, TechCrunch reports.

The financials disclosed in the filing paint a picture of a business that is moving beyond the “venture‑stage” growth phase. Cerebras reported $510 million in revenue for 2025, a figure that dwarfs the $1 billion‑plus annual revenues of many AI‑chip rivals, while posting a GAAP net income of $237.8 million. Adjusted for one‑time items, the company recorded a non‑GAAP net loss of $75.7 million, indicating that profitability is still emerging from heavy R&D and capital‑intensive production. Nonetheless, the revenue trajectory underscores the commercial traction of its wafer‑scale engine, which Feldman touts as “the fastest AI hardware for training and inference.” The filing therefore signals to investors that the firm is not merely a speculative play but a revenue‑generating entity with a clear path toward sustained margins.

Strategic partnerships are the linchpin of Cerebras’ market expansion and a key justification for the IPO. In the past quarter the startup secured an agreement with Amazon Web Services to deploy its chips in AWS data centers, a move that could give Cerebras a foothold in the cloud‑infrastructure ecosystem that currently favors Nvidia’s GPUs. More consequential, however, is the reported $10 billion‑plus deal with OpenAI, which, according to the Wall Street Journal, was motivated by OpenAI’s desire to preserve fast‑inference capability that Nvidia allegedly was reluctant to support. Feldman’s comment to the WSJ—“Obviously, [Nvidia] didn’t want to lose the fast inference business at OpenAI, and we took that from them”—highlights the competitive pressure the startup is exerting on the incumbent’s market share.

Analysts will likely scrutinize how Cerebras plans to scale production of its massive wafer‑scale chips, a process that historically has been fraught with yield challenges. The company’s ability to meet the demand generated by AWS and OpenAI contracts will be a decisive factor in translating its lofty valuation into tangible shareholder returns. Moreover, the timing of the offering—mid‑May, when the broader AI‑chip market is experiencing heightened volatility due to macro‑economic headwinds—adds a layer of risk. Yet the combination of robust cash reserves, a growing revenue base, and high‑profile customers positions Cerebras as a rare “growth‑at‑scale” candidate in a sector where most entrants are still pre‑revenue.

In sum, Cerebras’ IPO filing is less a flash‑in‑the‑pan fundraising event than a strategic inflection point. By locking in capital after two massive funding rounds, the company is signaling readiness to transition from rapid prototyping to mass production and broader market penetration. If the firm can sustain its revenue growth while navigating the manufacturing complexities of wafer‑scale silicon, it could reshape the competitive dynamics of AI hardware and offer investors a differentiated exposure to the fast‑evolving AI infrastructure landscape.