Cerebras IPO Roadshow Targets $3.5 Billion in Nvidia Challenge

AI chipmaker Cerebras Systems has launched its initial public offering roadshow, marking the company's second attempt to go public after withdrawing a previous filing last October. The Sunnyvale, California-based firm is targeting a share price between $115 and $125, which could generate up to $3.5 billion in proceeds, according to CNBC's reporting on the prospectus filing.

The company plans to list on the Nasdaq under the ticker symbol "CBRS." Accounting for the overallotment option, the offering could potentially raise as much as $4.03 billion, positioning Cerebras for a valuation around $26.6 billion at the high end of the range.

This isn't Cerebras' first rodeo with public markets. The firm initially sought to go public in 2024 but pulled the paperwork as its business model shifted from pure hardware sales toward operating a cloud service built on its proprietary chips. That pivot matters because it changes how investors should value the company—less like a traditional semiconductor manufacturer, more like a hybrid infrastructure play.

Financial performance has improved dramatically. Revenue climbed to $510 million for the year ending December 31, up from $290.3 million the prior year. More importantly, Cerebras posted earnings of $1.38 per share, a stark turnaround from a loss of $9.90 per share in the previous fiscal year. The fourth quarter alone showed $87.9 million in net income.

Co-founder and CEO Andrew Feldman is not selling shares in the IPO. He will retain 10.3 million shares post-offering, which could be worth up to $1.28 billion at the high end of the pricing range. This signals confidence—or at least a desire to maintain control—during a critical transition period.

The underwriting syndicate includes heavy hitters: Morgan Stanley, Citigroup, Barclays, and UBS. These banks are taking on significant risk given the current market environment. Relatively few technology companies have successfully gone public since central banks raised interest rates in 2022 to combat inflation, making investors more cautious about unprofitable names.

But the generative AI boom has created an exception. Money-losing competitor CoreWeave, which rents out Nvidia graphics processing units as a cloud service, raised $1.5 billion in its IPO last year. Investors are rabid about betting on companies that benefit from the AI trend, even if fundamentals are shaky.

Cerebras' competitive edge lies in its wafer-scale engine chips, designed to accelerate training and inference of large AI models. Unlike traditional chip manufacturing that cuts wafers into individual dies, Cerebras builds processors across the entire silicon wafer. The physical reality of this approach means fewer interconnect bottlenecks—data doesn't have to hop between chips as often. For developers, this translates to faster training cycles and less time waiting for compute clusters to finish their work (a problem that has plagued users for years, frankly).

The company's most significant validation came in January, when it announced a deal to provide up to 750 megawatts of AI computing power to OpenAI through 2028. The transaction is worth over $20 billion. This partnership directly challenges Nvidia's dominance in the AI infrastructure space, where the chip giant has held near-monopoly status for years.

Valuation comparisons are tricky. In February, Cerebras announced a venture round that gave it a $23 billion valuation, with Advanced Micro Devices among its investors. The IPO pricing suggests the market is willing to pay a premium for AI exposure, even if the company's path to sustained profitability remains unproven at scale.

There's also an option for underwriters to purchase an additional 4.2 million shares after the IPO, which would yield another $525 million in proceeds at the high end of the range. This green shoe provision gives the underwriters flexibility to stabilize the stock price in the early trading days.

The Azerbaijan news outlet News.Az reported on the IPO roadshow, citing Reuters as a source for the pricing details. While the story originated from U.S. financial markets, the coverage demonstrates how AI infrastructure developments are becoming global news, even in markets far removed from Silicon Valley.

Whether the market will sustain this valuation remains the real question. Cerebras is betting that its wafer-scale architecture can displace Nvidia's GPUs in enough data centers to justify the premium. The technology works in theory, but scaling production and convincing enterprise customers to switch from established Nvidia stacks is a different challenge entirely.

Time will tell if investors are buying into the AI narrative or the actual fundamentals. The IPO roadshow begins Monday, and the first real test comes when institutional investors start asking hard questions about margins, customer concentration, and whether that $20 billion OpenAI deal can actually materialize.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn