Cerebras AI growth story faces investor reality check

Cerebras Systems shares plunged about 16% in after-hours trading after the AI chipmaker's second-quarter revenue came in below Wall Street expectations, overshadowing stronger adjusted earnings and an upgraded full-year outlook. (Sources: Reuters, SiliconANGLE, Cerebras investor relations)

Cerebras reported $180 million in quarterly revenue, below the Street's estimate of about $194 million. The headline miss triggered concerns among investors about the company's ability to convert surging AI demand into sustainable, profitable growth.

The company's core revenue reached about $210 million, up roughly 50% year-on-year and ahead of Wall Street's $201 million expectation. Cerebras adjusts its core metric for pass-through revenue and the accounting impact of customer warrants, making the headline and core figures look notably different.

Cerebras posted a net loss of $450.5 million, compared with a profit of $309.5 million a year earlier. However, about $386.6 million of the loss was tied to stock-based compensation, highlighting why investors are focusing more closely on the company's underlying operating performance.

Despite the revenue miss, Cerebras raised its 2026 core revenue forecast to $880 million-$890 million, from its previous estimate of $855 million-$865 million. It also expects third-quarter core revenue of $214 million-$216 million, suggesting management remains confident about demand.

Cerebras is positioning its wafer-scale chips as an alternative to Nvidia's GPUs, particularly for low-latency AI inference. Its technology puts computing cores and memory on a single large silicon wafer, allowing certain AI workloads to run at high speed.

Cerebras' cloud and services business has expanded rapidly, with cloud revenue reaching about $126 million and growing several-fold from a year earlier. The shift toward recurring AI infrastructure and inference services could become increasingly important as hardware sales fluctuate.

Despite the strong AI demand narrative, Cerebras' order backlog remained around $25 billion during the quarter rather than increasing. Investors are therefore watching how quickly the company can translate that backlog into revenue and margins.

OpenAI is among Cerebras' biggest customers, with a major multi-year agreement supporting AI inference capacity. The company has also partnered with AMD, combining Cerebras chips with AMD's rack-scale architecture for faster inference workloads.

The latest sell-off shows investors are demanding more than rapid AI revenue growth. Cerebras now needs to demonstrate that it can scale production, expand margins and convert its massive backlog into profitable revenue while competing in a market dominated by Nvidia.