Three days after suggesting it could target between $4 billion and $40 billion, the AI chip maker has traded 28 million shares at between $115 and $125 each, aligning the deal with its February private valuation rather than above it.
Cerebras Systems was preparing to seek up to $4 billion at a valuation of approximately $40 billion in its long-awaited initial public offering. On Monday, the company filed an updated prospectus that priced the deal more cautiously: 28 million shares traded in a range of $115 to $125 each, raising up to $3.5 billion at the high end.
In that range, the implied valuation is closer to $26.6 billion, well below the figure circulated last week and roughly in line with the $23 billion private label Cerebras had after its Series H in February.
It is, by any reading, a more grounded number. Whether it is sufficiently substantiated is the question that the order book will answer.
By the updated brochure presented on MondayCerebras intends to list on the Nasdaq Global Select Market under the symbol CBRS, with Morgan Stanley, Citigroup, Barclays and UBS Investment Bank as joint book-running managers.
The 28 million share offering carries an option to subscribe for an additional 4.2 million shares, which would increase gross proceeds by approximately $525 million at the top of the range. CNBC notes that CEO Andrew Feldman, the company’s co-founder, will not sell shares in the IPO; He will own 10.3 million shares after the deal, worth up to $1.28 billion at the high end.
The financial profile underlying the offering is, by AI hardware standards, well developed. Cerebras’ most recent quarter showed revenue increased about 76 percent year-over-year to $510 million, with $87.9 million in net income.
The company’s framework relationship agreement with OpenAI, signed in January, promises up to 750 megawatts of inference capacity through 2028 in a contract that Cerebras has previously valued at more than $20 billion over its lifespan. That deal, more than any wafer-scale technology demonstration, has anchored the IPO narrative ever since.
Why did the rating go down?
The shift from a potential $4 billion ($40 billion) raise to a marketed $3.5 billion ($26.6 billion) raise is not dramatic, by IPO standards. It’s nothing either. There are three plausible readings.
The first is execution. Bidding over a broader range and a more sober valuation reduces the risk of a book being undersubscribed on launch day and gives the deal room to raise the price if demand is strong. After two years in which AI hardware listings have been priced below initial ranges as often as above, that discipline makes sense.
The second is the market context. Multiple AI software have been compressing through April, with Palantir down roughly 30 percent year-to-date and Citi cutting price targets across the AI platform space.
While Cerebras is more hardware than software, hardware listings have not been immune to that broader recalibration. Pricing within the now-public benchmarks for memory and computing is, again, sensible.
The third is concentration risk. Public market investors, in their diligence on the prospectus, will have noted that OpenAI is the dominant customer in the deal. The economics, if OpenAI executes as planned, are formidable.
If OpenAI’s computing demand recalibrates, even modestly, Cerebras’ near-term revenue trajectory recalibrates with it. That risk does not disappear at $26.6 billion; However, it is easier to subscribe with that figure than with $40 billion.
What would it mean if it listed?
If Cerebras lists at the top end of the range, the offering would be one of the largest U.S. tech IPOs in 2026, ahead of any new entrant in AI silicon since Nvidia revived its public market presence three years ago.
Reuters and IFR have flagged the deal as an increase of approximately $3 billion. in terms of Nasdaq IPO; The exact number will depend on whether the insurer’s over-allotment is exercised and the final price. Either way, this is the largest first-day public equity event to date on dedicated AI training and inference hardware.
It would also be a public test of a thesis that Cerebras has carefully built for a decade: that its wafer-scale chip, an architecturally radical alternative to packing many smaller GPUs on a board, is now commercially defensible at scale. The OpenAI contract is the most visible vote of confidence in that thesis. The IPO, if the price is in line, is the second.
The path from presentation to first transaction is not a formality. Cerebras’ previous IPO attempt, in 2024, failed after the US Committee on Foreign Investment opened a review of G42’s stake in the company; The eventual restructuring of that stake into non-voting stock paved the way for the current attempt.
Public market investors will have read that risk factors section carefully. They will also have noted Cerebras’ customer concentration, the extent to which margin will depend on the cost of advanced packaging, and the strategic question of whether Nvidia’s next-generation accelerators erode the use cases that Cerebras’ wafer-scale design currently best serves.
None of those are fatal. However, they are the substance behind the lower valuation. Cerebras has presented a serious prospect in a market that, right now, takes AI hardware more seriously than AI software. The rest depends on the order book.
