What is the Cerebras IPO roadshow? The Cerebras IPO roadshow is the formal investor marketing campaign Cerebras Systems launched on May 4, 2026, ahead of pricing its initial public offering on Nasdaq under the ticker CBRS. The company is offering 28 million shares at $115 to $125 each to raise approximately $3.5 billion at a fully diluted valuation of up to $26.6 billion, with pricing expected around May 13, 2026. For CRE data center investors, this listing is the cleanest public market read on AI compute demand outside of Nvidia, landing as hyperscaler capex commitments reshape the data center asset class. For broader context, see our complete guide on the best AI tools for commercial real estate investors.
Key Takeaways
- Cerebras priced its IPO range at $115 to $125 per share for 28 million shares, targeting a $3.5 billion raise and a $26.6 billion fully diluted valuation on Nasdaq under ticker CBRS.
- The $26.6 billion target is a meaningful step down from the $35 billion valuation Cerebras pursued in its April 17, 2026 prospectus, signaling investor caution on AI chip multiples.
- OpenAI has committed more than $20 billion to Cerebras compute over multiple years, equivalent to 750 megawatts of dedicated AI capacity, a number that translates directly into data center demand.
- Amazon Web Services signed a binding term sheet in March 2026 to become the first hyperscaler to deploy Cerebras systems, broadening the footprint beyond OpenAI.
- Cerebras reported $510 million in fiscal 2025 revenue, up from $290.3 million, with a swing to a profit of $1.38 per share from a loss of $9.90 per share the prior year.
- For CRE investors, the 750 MW OpenAI commitment alone equals roughly 5 to 7 hyperscale data center campuses worth of forward power, land, and grid demand.
Cerebras IPO Roadshow Terms in Detail
According to the updated S-1 prospectus filed with the SEC on May 4, 2026, Cerebras is offering 28 million Class A shares at a price band of $115 to $125. At the high end, the offering raises $3.5 billion in gross proceeds with a fully diluted valuation of approximately $26.6 billion. Underwriters Morgan Stanley, Citigroup, Barclays, and UBS hold a 30 day overallotment option for an additional 4.2 million shares, which would lift proceeds by roughly $525 million if exercised. According to HPCwire reporting on the roadshow launch, pricing is expected around May 13, 2026, with shares listing on the Nasdaq Global Select Market shortly thereafter under ticker CBRS.
The roadshow valuation is a measured retreat from the $35 billion target floated in the April 17, 2026 filing, which we covered in detail in Cerebras files for $35 billion IPO as OpenAI chip deal doubles to $20 billion. The reset still represents a 16% premium to the $23 billion private valuation Tiger Global led in January 2026 and more than threefold the $8.1 billion valuation Cerebras carried in September 2025.
Why the Valuation Cut Matters for CRE
The markdown from $35 billion to $26.6 billion is the most CRE relevant signal in the prospectus. Public markets are repricing AI infrastructure equity even as private markets continue to set fresh highs. Data center valuations, lease economics, and hyperscaler tenant credit increasingly track the cost of capital for the AI chip players who anchor demand. When a critical compute supplier accepts a 24% valuation cut to access public capital, it tightens financing conditions for the entire AI build out and forces hyperscalers to discipline site selection.
Despite the cut, $26.6 billion on $510 million of trailing revenue still implies more than 50 times sales, a multiple pricing in continued exponential growth in AI training and inference workloads. This is the same demand curve driving $145 billion of Meta capex and $185 billion at Google. See our coverage of Big Tech's $700B AI capex surge in 2026.
The 750 MW OpenAI Deal Translated to Data Center Reality
The single most important customer disclosure in the Cerebras prospectus is the multi year contract with OpenAI valued at more than $20 billion. Under that deal, OpenAI will absorb 750 megawatts of Cerebras compute capacity over the contract term. For CRE underwriting purposes, 750 megawatts is enormous. A typical hyperscale data center campus draws between 100 and 250 megawatts at full build out. Translating the OpenAI commitment into physical real estate yields the equivalent of roughly 5 to 7 hyperscale campuses worth of power demand sitting behind a single chip vendor relationship.
This concentration cuts two ways for CRE investors. On the upside, it converts a public market chip investment into a quantifiable forward pipeline of land, power, and shell capacity needs. On the downside, it ties data center underwriting to OpenAI's ability to fund and deploy that compute. Our analysis of OpenAI's compute commitments and CFO Friar's $600B warning walks through that counterparty risk.
AWS as Hyperscaler Validation
In March 2026, Amazon Web Services signed a binding term sheet to become the first hyperscaler to deploy Cerebras systems inside its own data centers. AWS adoption converts Cerebras from a single customer dependent business into a credible third party silicon option for the broader hyperscaler stack. For CRE investors with data center exposure, hyperscaler tenants drive AI lease demand, and a non Nvidia chip in the AWS catalog increases the menu of workloads that can land in any given facility. AWS expansion was a centerpiece of Amazon's Q1 2026 earnings and $44 billion capex commitment.
How CRE Investors Should Read the IPO
Three actions follow from the roadshow disclosures. First, treat the $26.6 billion valuation as the new public benchmark for AI chip equity through 2026, which compresses comparable multiples for any private chip platform pitched as a data center anchor. Second, map the 750 MW OpenAI commitment to your portfolio's sub markets, especially Texas and Ohio where Cerebras and OpenAI have telegraphed deployment intent. Third, watch the underwriter overallotment exercise as a real time signal of secondary demand for AI infrastructure paper. The team at The AI Consulting Network helps CRE investors translate these signals into deal level underwriting assumptions.
Power, Land, and Grid Implications
The Cerebras prospectus reinforces what CBRE's Global Data Center Trends and JLL outlook reports have flagged: power availability has displaced location as the binding constraint on AI data center supply. Cerebras WSE-3 systems run at extreme density, with single rack power draws beyond 50 kilowatts that require liquid cooling at the chassis level. Owners of pre 2020 single tenant stock face a choice between expensive retrofits or repositioning toward inference and edge workloads where legacy power densities still work. Repositioning cap rates currently sit between 7.5% and 9.0% for converted shells, compared to 5.5% to 6.5% for purpose built AI ready capacity in markets like Phoenix and Northern Virginia. Avi Hacker, J.D. at The AI Consulting Network helps owners stress test repositioning economics against tenant demand.
Risks Disclosed in the Prospectus
The S-1 highlights four risk vectors CRE investors should track. Customer concentration remains acute even after AWS, with OpenAI representing a majority of contracted revenue. Foundry concentration is total, since only Taiwan Semiconductor Manufacturing Company can fabricate the wafer scale dies. Regulatory risk persists from the prior G42 review by the Committee on Foreign Investment in the United States, even though G42 has restructured into non voting shares. Competitive risk from Nvidia's next generation accelerators could erode the inference latency advantage that drives Cerebras pricing power. Each of these risks affects the durability of the 750 MW demand pipeline being priced into AI lease assumptions.
Frequently Asked Questions
Q: When does the Cerebras IPO price?
A: The Cerebras IPO roadshow launched on May 4, 2026, with pricing expected around May 13, 2026. Shares will list on the Nasdaq Global Select Market under ticker CBRS shortly after pricing.
Q: What is the Cerebras IPO valuation?
A: At the high end of the $115 to $125 share price band, Cerebras is targeting a fully diluted valuation of approximately $26.6 billion on $3.5 billion of gross proceeds from 28 million shares.
Q: How does the Cerebras IPO affect CRE data center demand?
A: Cerebras has more than $20 billion in committed compute contracts with OpenAI representing 750 megawatts of capacity, which translates into roughly 5 to 7 hyperscale data center campuses of forward power and shell demand. The AWS binding term sheet from March 2026 broadens this demand across hyperscaler markets.
Q: Why did Cerebras cut its IPO valuation from $35 billion to $26.6 billion?
A: The downward revision reflects investor pushback on AI chip multiples, broader IPO market softness in early 2026, and customer concentration disclosures. Even at the reduced level, the IPO implies more than 50 times trailing revenue.
Q: What is the Cerebras Wafer Scale Engine 3?
A: The WSE-3 is Cerebras's flagship AI processor, fabricated by TSMC at chip scale dimensions roughly 58 times larger than a leading GPU. The company reports 15 times faster inference and 2,625 times more memory bandwidth than Nvidia's B200 package.