Cerebras IPO Ignites AI Chip Wars with $6.38B War Chest

SUNNYVALE, CA – May 15, 2026 – Cerebras Systems, the ambitious builder of the world's largest computer chips, has officially entered the public market with a resounding boom. The company announced today the closing of its initial public offering, raising a staggering $6.38 billion in gross proceeds and signaling a new, high-stakes chapter in the battle for AI hardware supremacy.

The offering, which included the full exercise of the underwriters' option to purchase additional shares, saw 34,500,000 shares of Class A common stock sold at $185.00 per share. The successful closing provides Cerebras with a massive capital injection to scale its operations and intensify its challenge against industry titan NVIDIA.

A Blockbuster Debut in the AI Gold Rush

Investor appetite for a piece of the burgeoning AI infrastructure market proved voracious. Cerebras's stock, trading under the ticker “CBRS” on the Nasdaq Global Select Market, made a spectacular debut on May 14. After being priced at $185, shares opened for trading at $350 and closed the day up 68% at $311.07.

The debut stands as the largest U.S. tech IPO since 2019 and the largest of 2026 thus far. The intense demand was evident long before the first trade, with reports indicating the offering was oversubscribed by more than 20 times. This fervor led underwriters, including lead managers Morgan Stanley and Citigroup, to raise the IPO price multiple times from its initial range.

This explosive market reception values Cerebras at a staggering multiple of its 2025 revenue of $510 million. While the company did report a net income of $238 million last year, a significant portion was attributed to a one-time accounting gain. The high valuation signals that investors are betting heavily on the company's future growth and its potential to capture a significant share of an AI hardware market projected to surpass $900 billion by 2035. However, some market observers note the high price-to-sales ratio and the significant customer concentration—with two entities in the UAE accounting for 86% of 2025 revenue—as potential risks for investors to watch.

Challenging the Throne: A Wafer-Scale Revolution

The billions raised are not just for show; they are fuel for a direct technological assault on the status quo. Cerebras's core innovation and primary weapon is the Wafer-Scale Engine 3 (WSE-3), a processor of unprecedented scale. Instead of dicing a silicon wafer into hundreds of small chips, Cerebras uses the entire wafer as a single, massive processor.

The WSE-3 is a marvel of engineering, packing 4 trillion transistors and 900,000 AI-optimized cores onto a single piece of silicon 58 times larger than a leading GPU. This design is fundamentally different from NVIDIA's approach, which involves linking thousands of smaller GPUs together in vast clusters. By keeping all compute and memory on one chip, Cerebras aims to eliminate the communication bottlenecks and latency issues that can slow down large-scale AI models.

With 44GB of on-chip SRAM, the WSE-3 boasts a memory bandwidth of 21 petabytes per second, a figure thousands of times greater than top-tier GPUs. The company leverages this architectural advantage to claim its systems can deliver inference up to 15 times faster than competing solutions for leading AI models. In specific scientific computing workloads, research has indicated speedups of over 200 times compared to an NVIDIA H100. This performance comes with the promise of a lower total cost of ownership and greater power efficiency, critical factors for the operators of power-hungry data centers.

From Labs to Hyperscalers: Powering the AI Frontier

Cerebras has already translated its technological promise into significant commercial and research partnerships. The company's customer list includes government research institutions like Lawrence Livermore National Laboratory, healthcare leaders such as the Mayo Clinic, and pharmaceutical giants like GlaxoSmithKline, all using its systems to accelerate discovery.

Most significantly, Cerebras has secured landmark deals that position it as a key infrastructure provider for the generative AI boom. In January, the company signed a multi-year agreement reportedly worth over $10 billion with OpenAI to deploy hundreds of megawatts of inference capacity. This was followed by a binding term sheet with Amazon Web Services (AWS) in March to integrate Cerebras systems into its cloud data centers.

Furthermore, its partnership with the UAE-based technology group G42 is building out the “Condor Galaxy,” a network of nine interconnected AI supercomputers. This initiative is a cornerstone of the growing global push for “sovereign AI,” where nations build their own AI infrastructure to ensure technological independence and data security—a market projected to grow at over 30% annually.

These high-profile collaborations demonstrate the real-world applicability of Cerebras's technology for the most demanding AI workloads, from training frontier models to enabling real-time enterprise applications like advanced search and AI-powered coding agents.

A Founder-Led Moonshot with Billions in Fuel

The company's journey began in 2015 as a “moonshot” project by a team of five founders, led by CEO Andrew Feldman, who had previously sold their microserver company SeaMicro to AMD. Their vision to build a wafer-scale chip, an idea long explored but deemed too difficult by many in the industry, has now been validated by the public markets. The founding team retains tight control over their creation, holding Class B shares that give them approximately 99.2% of the total voting power post-IPO.

With the $6.38 billion raised, this founder-led team now has the capital to dramatically scale production of its CS-3 systems, deepen its R&D into next-generation processors, and expand its global sales and support infrastructure. The IPO transforms Cerebras from a well-funded startup into a publicly traded powerhouse, armed with the resources necessary to compete at the highest level. As enterprises and governments race to deploy AI, Cerebras is now better positioned than ever to prove that its radical approach to chip design is not just an alternative, but the future of high-performance computing.