Cerebras seeks up to $4 billion in IPO, targets valuation of about $40 billion

Cerebras is returning to the public markets and aiming for one of the most notable technology deals of the year. The AI chip manufacturer wants to raise up to $4 billion at a valuation of around $40 billion — following a failed IPO attempt in 2024 and against the backdrop of a major contract with OpenAI.

Return to the

Markets Back in autumn 2024, Cerebras filed documents for an IPO, but the market entry was delayed by scrutiny from American regulators. The problem turned out to be not the technology, but the structure of shareholders and clients: G42 from Abu Dhabi played an important role in the company's business, and this very connection raised questions from the Committee on Foreign Investment in the United States. In October 2025, the registration was withdrawn, and the story looked like an example of how geopolitics could stop even a very prominent AI startup.

Now the company has returned with a more aggressive plan. On April 17, 2026, Cerebras resubmitted its S-1 to the SEC, and on May 2, it became known that the company expects to raise up to $4 billion at a valuation of approximately $40 billion. This is noticeably higher than the range analysts had previously expected, and nearly five times higher than the private valuation of $8.

1 billion at which Cerebras raised money in September 2025. If the placement goes through at the upper end, it could become the first major AI hardware IPO in 2026. The company plans to list on Nasdaq under the ticker CBRS.

The placement is being arranged by Morgan Stanley, Citigroup, Barclays, and UBS. The composition of these banks itself shows that Cerebras no longer looks like a niche experiment: it is a bid for a major public story, where the company will be compared not with other startups, but with leaders of the AI infrastructure market and providers of computing power with very high expectations.

The

Bet on Inference Cerebras's main argument to investors is not simply an unusual chip architecture, but the presence of an anchor client. In the updated prospectus, the company revealed a multi-year agreement with OpenAI for inference capacity through 2028. For the market, this is an important signal: the startup is not selling the promise of someday finding demand, but has already embedded itself in the real supply chain for computing for one of the largest players in the AI market, and this changes the tone of the entire deal.

Against this backdrop, even current financial metrics begin to look different. Cerebras's revenue for 2025 amounted to $510 million, which is 76% more than a year earlier. For a hardware manufacturer, this is a notable growth rate, but more importantly, the contract with OpenAI provides visibility for future utilization and helps explain the jump in valuation.

Infrastructure investors currently particularly value companies that have a large, clear, and long-term buyer of capacity. up to $4 billion the company wants to raise in the IPO approximately $40 billion — target valuation of the business $510 million in revenue Cerebras received in 2025 up to 750 MW of inference capacity planned to be deployed for OpenAI by 2028 * the company's valuation grew from $8.1 billion in September 2025 to $23 billion in February 2026 It is also important that Cerebras is betting not on a direct attack on Nvidia in model training, but on a narrower and rapidly growing segment — inference.

That is where ready-made models are deployed, user requests are served, and there is particularly strong pressure on costs. Cerebras's logic is simple: if its wafer-scale processors really provide advantages in response speed and compute density, it can occupy its own niche alongside Nvidia, rather than instead of it.

Where the Risks Remain Despite the new momentum, the deal has several weak points.

First, the G42 issue has not completely disappeared. In March 2025, the company received regulatory approval after the ownership structure was changed and G42's stake was converted to non-voting shares. Formally, the obstacle has been removed, but for some public investors, the very fact of previous dependence on a single foreign client and shareholder will still remain a risk factor.

Second, Cerebras's technological distinction is simultaneously its vulnerability. The company is known for its wafer-scale chips — processors so large that they are usually compared not with conventional GPUs, but with an entire silicon wafer. This sets Cerebras apart from Nvidia, but such an approach is more difficult to scale and manufacture in volume.

The public market usually only readily pays for uniqueness when it sees clear manufacturing economics and margins over the long term. Finally, there is a broader market question. In late 2025, several major AI listings came to market weaker than expected, and investors became more cautious about unprofitable infrastructure companies with very high private valuations.

Therefore, Cerebras's IPO will be a test not only for itself, but also for the entire thesis that the market is ready to automatically transfer the buzz around AI into the value of hardware.

What

This Means If Cerebras is indeed able to list on terms close to those announced, it will be an important signal for the entire AI infrastructure segment. The market will show that it is willing to generously value not only creators of models, but also providers of computing power with strong contracts and clear specialization. If demand turns out to be weaker, it will be a reminder: even in AI, public money already requires not just flashy technology, but solid proof of economics.