Cerebras IPO Doubles on Day One: What the Biggest AI Chip Debut Means for Investors

Cerebras Systems began trading on the Nasdaq Global Select Market on Wednesday under the ticker CBRS, and the most anticipated AI chip IPO of the year did not disappoint. Shares opened at approximately $350, nearly double the $185 IPO price, giving the company an implied valuation approaching $80 billion before a single trade had cleared as a public company.

The company raised at least $5.55 billion from the sale of 30 million Class A shares. The IPO was 20 times oversubscribed. Arm Holdings and SoftBank reportedly approached Cerebras with a preliminary acquisition offer before the offering. The company proceeded with the IPO anyway.

If you are trying to understand what just happened and whether it matters beyond Wall Street, here is the breakdown.

## What Cerebras Actually Does

Cerebras is not trying to be the next Nvidia. It is trying to own the specific part of AI computing where Nvidia's GPU architecture has inherent limitations: inference speed and latency at the extreme end of performance requirements.

The company's flagship product, the Wafer-Scale Engine 3, is physically different from any chip Nvidia or any other GPU manufacturer produces. Where conventional chips are the size of a fingernail, the WSE-3 encompasses an entire silicon wafer. CEO Andrew Feldman put it simply: "We built a chip the size of a dinner plate. It's 58 times larger than any chip previously built. In AI, bigger chips are faster."

The result is a chip with 4 trillion transistors and 900,000 AI-optimized compute cores that can train and run AI models faster than clusters of conventional GPUs. The tradeoff is that the chip is expensive to manufacture, requires specialized cooling, and only makes economic sense for organizations running the largest AI models at scale.

## The Financials: Profitable and Growing

This is what separates Cerebras from most high-profile IPOs in recent memory: it makes money. Cerebras generated $510 million in revenue and $238 million in net profit in 2025. That is a profit margin of roughly 47 percent, which is extraordinary for a hardware company at this stage.

The company also has a $20 billion multi-year contract with OpenAI, which provides revenue visibility that most chip companies would envy. OpenAI is the single largest customer, and while customer concentration is a risk, it is hard to imagine a better anchor client in the current market.

## Why the Stock Doubled

Several factors drove the first-day pop. First, AI infrastructure spending shows no signs of slowing. Microsoft, Amazon, Google, and Meta are collectively projected to spend over $300 billion on data center capital expenditures in 2026, and a significant portion of that goes to AI chip procurement. Any company that can credibly claim to offer an alternative to Nvidia's monopoly is going to attract investor interest.

Second, the supply-demand dynamics of the IPO itself favored a big first day. Only 30 million shares were offered, the IPO was 20 times oversubscribed, and institutional investors who wanted exposure to the AI chip space outside of Nvidia had limited options. When demand vastly exceeds supply, the price goes up.

Third, Cerebras is profitable, which makes it a legitimate alternative to Nvidia rather than a speculative bet on future growth. The market is rewarding companies that can demonstrate both revenue and earnings in the AI space, and Cerebras checks both boxes.

## The Risks Nobody Is Talking About

Every IPO has risks, and the ones that double on day one tend to have risks that get overlooked in the excitement.

Customer concentration is the most obvious. OpenAI represents a dominant share of Cerebras' revenue. If OpenAI's needs change, if they develop their own silicon, or if they shift spending to a competitor, Cerebras' revenue would be materially affected. The $20 billion contract provides visibility, but contracts can be renegotiated, and the AI landscape changes faster than most.

Manufacturing risk is real. The WSE-3 is produced on TSMC's advanced nodes, and any disruption to TSMC's capacity, whether from geopolitical tensions with China, natural disasters in Taiwan, or simple production bottlenecks, could delay shipments and erode Cerebras' competitive position.

Valuation risk is inherent at these levels. An $80 billion market capitalization for a company with $510 million in revenue implies a price-to-sales ratio of over 150. Even by AI standards, that is aggressive. Nvidia trades at roughly 25 times sales, and it has a 90 percent market share. Cerebras is priced for perfection, and anything short of flawless execution will disappoint.

## What This Means For You

If you are a retail investor, the Cerebras IPO is a reminder that the AI investment landscape is broadening beyond Nvidia. For the past two years, buying Nvidia has been the default AI trade. Cerebras' successful debut shows that the market is ready for alternatives, and that is healthy for competition and for investors who want diversification.

If you work in tech, Cerebras' wafer-scale approach validates the thesis that the industry is moving away from the GPU-centric model of AI computing. Expect more companies to explore alternative architectures, whether wafer-scale, photonic, or neuromorphic. The GPU is not going away, but it is no longer the only game in town.

If you are watching from the sidelines, the lesson is simpler: AI infrastructure spending is still accelerating, and the companies that enable it, whether Nvidia, Cerebras, or the next entrant, are going to capture an outsized share of value. The question is not whether AI will reshape computing, but who will profit most from the reshaping.