There is something almost poetic about watching a stock that opens at $350 – nearly double its IPO price, pull back 10% two days later. The market giveth. The market taketh. And the people who got in at $185 are still smiling while everyone who chased the opening print is doing the math.
Cerebras Systems (CBRS) had one of the most dramatic IPO debuts in years. Priced at $185 on Wednesday evening, May 13, above its already-raised range of $150 to $160, CBRS opened Thursday May 14 at $350 and surged to $385 before a trading halt intervened. The shares closed on day one at $311 – a 68% first-day gain. By Friday’s close, CBRS was at $279.72, down another 10.08%.
The AI chipmaker raised approximately $6.38 billion in gross proceeds after underwriters exercised their full overallotment option, according to the company’s May 15 closing announcement. At peak pricing, Cerebras was valued at roughly 130 times trailing revenue. That number deserves attention because it explains everything about what happened next.
What Cerebras actually is, and why the market went wild for it
Cerebras is not building another GPU. It is building something fundamentally different. The company’s flagship Wafer-Scale Engine 3 is, according to Cerebras, the largest commercialized AI processor in the world. In fact, it is 58 times larger than a leading GPU chip.
It delivers AI inference up to 15 times faster than GPU-based solutions while using a fraction of the power per unit compute. The company pitches this as the answer to Nvidia fatigue – a faster, more efficient path to running AI workloads at scale.
More AI:
- Micron sits at the center of a red-hot chip rally
- IBM CEO sends blunt message on AI and quantum computing
- Anthropic CEO makes shocking admission about AI
The customer roster adds credibility: OpenAI, Amazon Web Services, Meta Platforms, and IBM, according to the company’s prospectus. The January deal with OpenAI – a multi-year agreement covering 750 megawatts of inference capacity, expandable to 2 gigawatts by 2030 and described as worth more than $20 billion at full expansion – was the catalyst that turned investor interest into investor frenzy. AWS signed a binding term sheet in March to deploy Cerebras systems directly inside its data centers.
Revenue momentum backed the enthusiasm. Cerebras grew revenue 76% in 2025 to $510 million, according to Forbes. The company swung to net income of $238 million, though much of that reflected a one-time accounting gain – and operating losses are still widening.
The valuation math is making investors nervous on day two
Here is where the story gets complicated. The Wall Street Journal analysis noted that Cerebras’s closing price on day one valued the company at approximately 134 times revenue for the past four quarters. For context, Nvidia, which generated revenue of $62.3 billion, up 22% from Q3 and up 75% from a year ago, trades at roughly one-fifth that revenue multiple.
My review of the concentration risk inside the business adds another layer of concern. Approximately 86% of Cerebras’s 2025 revenue came from just two UAE-linked customers, The Motley Fool confirmed. That is an extraordinary degree of customer concentration for a company being valued at tens of billions of dollars.
Related: Cerebras IPO to land amid chipmaker mania on Wall Street
The historical data on IPOs of this magnitude is sobering. Research from University of Florida finance professor Jay Ritter, tracking IPO returns from 1980 through 2024, shows newly public companies underperforming comparable firms by an average of 3.6% annually in their first five years.
For IPOs since 2010, the first-year gap versus comparable non-IPO firms has been approximately 9 percentage points. Snowflake, the last IPO of comparable scale, opened at $245 in 2020 and remains underwater for investors who bought at that first-day close, despite massive revenue growth in the years since.
What Wedbush’s Dan Ives says about where AI IPOs go from here
Not everyone is hitting the brakes. Wedbush analyst Dan Ives, speaking on Bloomberg after Cerebras’s debut, called the IPO a watershed moment for the tech sector and pushed back hard against any bearish read on the AI cycle. Ives described the current AI buildout as only in “the third inning” of a much longer story, citing a 20% quarter-over-quarter acceleration in AI demand from his recent Asia visits.
Ives also described Cerebras as the appetizer, with future IPOs from companies like SpaceX, OpenAI, and Anthropic potentially generating even larger market enthusiasm. He argued the Nasdaq could reach 30,000 within six to nine months as AI infrastructure spending continues to accelerate.
More Wall Street:
For investors watching CBRS from the sidelines, the Ives framework and the Ritter data represent the two poles of the debate. One says this is the early innings of a historic cycle, and Cerebras has the technology and contracts to grow into its valuation. The other says IPO buyers at elevated prices have historically regretted the rush.
Well, both can be right about the technology and wrong about the entry price, and that tension is what makes CBRS one of the most interesting and most dangerous new stocks in the market right now.
Related: Wedbush has blunt message for SoundHound AI stock investors
