Cerebras IPO Price Target Raised to $160 Per Share

Cerebras Systems Prepares to Raise IPO Price Target Amid Record Investor Demand

Artificial intelligence chipmaker Cerebras Systems Inc. is preparing to significantly raise its IPO price range. The Sunnyvale, California-based company now targets a price between $150 and $160 per share. That compares to its original target range of $115 to $125 per share. The revised figures reflect explosive investor appetite for AI-focused technology companies.

Two anonymous sources familiar with the matter shared the details with Reuters. The company has not yet made the information public through official channels. Cerebras did not immediately respond to a request for comment. All figures remain subject to change before the official pricing date.

The company also plans to increase the number of marketed shares. It raised that figure from 28 million to 30 million shares. At the top of the new price range, Cerebras expects to raise roughly $4.8 billion. That marks a sharp increase from the original target of $3.5 billion under earlier terms.

Demand Surges More Than 20 Times Available Supply

Investor enthusiasm for Cerebras shares has far exceeded expectations. The IPO drew orders for more than 20 times the number of shares available. That level of oversubscription reflects intense market interest in high-performance AI chip providers. Cerebras now manages surging demand ahead of its May 13 pricing date.

The AI sector has experienced a broad and rapid expansion in recent years. Demand for specialized chips now outpaces available supply across the industry. Semiconductors have become a critical bottleneck across the entire technology supply chain. Companies like Cerebras stand at the center of that constraint.

Cerebras competes directly with Nvidia in the specialized chip market. Nvidia currently dominates the sector with its Blackwell B200 graphics processing units. Cerebras differentiates itself with a fundamentally different chip architecture. Its approach focuses on wafer-scale integration rather than traditional GPU design.

The WSE-3 Chip Sets Cerebras Apart From Rivals

Cerebras produces wafer-sized AI chips known as the WSE-3. These chips are several times larger than Nvidia’s Blackwell B200 graphics processing units. The WSE-3 features a 44-gigabyte pool of SRAM memory. SRAM carries significantly more transistors per square millimeter than standard server DRAM.

That architectural advantage makes SRAM considerably faster than conventional memory types. It is also several orders of magnitude more expensive. Each processor includes a quartz crystal that moves millions of times per second. Every movement corresponds to one clock cycle, the basic unit of computing time.

Cerebras states that the WSE-3’s 900,000 cores access the onboard SRAM pool with one clock cycle of latency. That speed far exceeds what standard graphics cards can deliver. The result translates directly into faster inference performance. Inference refers to the process by which an AI model responds to user queries.

The CS-3 Appliance Powers Real-World Deployments

Cerebras ships the WSE-3 as part of a physical appliance called the CS-3. That appliance weighs 1.8 tons and houses the chip within an enterprise-ready system. Customers can link multiple CS-3 units together into larger clusters. Cerebras also sells auxiliary devices that enable those cluster configurations.

AI labs are actively shifting focus from model training to model deployment. That shift plays directly to Cerebras’ core strengths in inference computing. Its chips perform especially well on inference tasks. Traditional GPU chips have long dominated model training but face limitations in deployment contexts.

The company targets AI labs and enterprises running large-scale AI models. Its chips offer distinct advantages when AI systems must respond to high volumes of user queries. That positions Cerebras well in a market that increasingly demands faster and more efficient inference. Demand for inference-optimized hardware continues to accelerate globally.

Cerebras Revives IPO After Earlier Regulatory Setbacks

This upcoming IPO marks Cerebras’ second attempt to go public. The company first filed IPO paperwork in 2024. It later pulled that plan after facing regulatory complications. Its partnership with G42, a UAE-based AI company, triggered a national security review.

The Committee on Foreign Investment in the United States conducted that review. The committee eventually cleared the deal, and Cerebras proceeded with its plans. G42 had provided more than 80% of Cerebras’ revenue in the first half of 2024. That concentration of revenue from a foreign partner raised concerns among regulators.

Bloomberg News previously reported that Cerebras planned to raise its IPO price range to between $125 and $135 per share. The latest figures represent a further and more aggressive increase beyond that earlier revision. The final share count and price remain subject to change before May 13. Investors and analysts continue to watch the offering closely.

IPO Timing Aligns With Peak AI Investment Cycle

Cerebras has timed its public market debut strategically. It enters the market at a moment of peak institutional enthusiasm for AI infrastructure. The company joins a wave of technology firms capitalizing on surging AI investment. Market conditions currently favor high-growth, hardware-focused AI companies.

The broader AI chip market continues to expand at a rapid pace. Enterprises worldwide accelerate their adoption of AI tools and infrastructure. That adoption drives consistent and growing demand for specialized processing hardware. Cerebras positions itself as a direct beneficiary of that structural trend.

The official pricing of the Cerebras IPO takes place on May 13. The final price and total capital raised will depend on last-minute market conditions. Investors expect the offering to attract continued strong demand through pricing day. The IPO stands as one of the most anticipated technology market debuts of the year.