Cerebras

A $5.5B Nasdaq IPO in May 2026 pushed Cerebras from AI-inference challenger to a heavily scrutinized public chipmaker tied to OpenAI and AMD.

Who they are

Cerebras is an AI chip designer centered in this coverage on wafer-scale computing and AI inference. Its role has shifted from a challenger pitching an inference service against Nvidia to a public supplier whose technology is being deployed through OpenAI and integrated with AMD server infrastructure.

The recent arc

Coverage accelerated sharply in 2026, first around an OpenAI capacity agreement and then around Cerebras’s market debut. The January reports said OpenAI had struck a multibillion-dollar, three-year agreement for 750 MW of Cerebras computing capacity, while Reuters positioned Cerebras among alternatives OpenAI had explored amid dissatisfaction with some Nvidia inference chips. In May, Cerebras priced 30 million shares at $185, raising $5.5B, before shares closed 68% higher in its Nasdaq debut; stories emphasized both the scale of the offering and the resulting valuation.

Post-IPO coverage has turned from financing to execution and partnerships. June earnings paired 94% year-over-year Q1 revenue growth and a narrower loss with a warning of Q2 core-margin compression. In July, AMD said it was connecting Helios server racks with Cerebras wafer-scale chips for inference, and August reports showed Q2 revenue growing 74% to $180M but missing estimates even as annual revenue and gross-margin forecasts rose. The latest OpenAI Ultrafast API preview, powered by Cerebras, brought product-performance claims back to the center of the story.

The tension

The central tension is whether Cerebras can convert demand for alternatives to Nvidia into durable, profitable inference infrastructure. OpenAI’s use of Cerebras gives the company a high-profile route to deployment, while the AMD integration suggests a complementary systems approach rather than a standalone-chip contest. But the earnings coverage shows investor sensitivity to revenue expectations and margins, even alongside rapid growth and stronger full-year guidance.

Why it matters

Cerebras is becoming a test case for whether specialized AI-inference hardware can gain strategic relevance beside GPU incumbents through major customer platforms and heterogeneous deployments. If OpenAI’s Ultrafast tier and the AMD collaboration translate into sustained usage, they could validate that route; whether growth can support margins and public-market expectations remains the unresolved constraint reflected in recent coverage.

Related: OpenAI · Nvidia · AMD is working with Cerebras to connect AMD's server racks to Cerebras · Cerebras shares closed up 68% at $311.07 in the company's Nasdaq debut · IPO