Cerebras Shares Jump 17% After Hours as It Raises 2026 Forecast
Cerebras Systems lifted its 2026 revenue forecast to $940 million–$960 million and signed Microsoft as a cloud customer on November 12, sending shares up more than 17% in extended trading as investors reassessed AI spending worries.
Cerebras Systems shares rose more than 17% in extended trading on Thursday, November 12, after the chip and cloud company raised its 2026 revenue forecast and said it had signed Microsoft as a customer.
The company now expects adjusted revenue of $940 million to $960 million this year. It had previously guided to $880 million to $890 million.
Third-quarter results
Adjusted sales rose 70% to $230.9 million in the third quarter, beating the average analyst estimate of $214.70 million. Adjusted cloud and services revenue, which reflects the OpenAI deal, more than tripled to $144.9 million.
Cerebras posted an adjusted profit of $29.82 million. A year earlier it reported an adjusted loss of $10.86 million, and analysts had expected a loss of $38.4 million, according to data compiled by LSEG.
Why investors were nervous
The stock had fallen 47.2% from its May IPO price. Investors have grown wary of the huge infrastructure spending needed to support AI growth, and the outlook had also been clouded by US voter opposition to data centers and uncertainty over Jensen Huang's exit from Nvidia.
The new forecast pointed to lasting demand for AI computing power, despite those concerns.
OpenAI, Microsoft and capacity limits
Cerebras is racing to support a computing-power deal with OpenAI. That deal underpins remaining performance obligations of $26.1 billion as of September 30. That figure measures contracted work not yet delivered.
CEO Andrew Feldman said on a call with analysts that the Microsoft cloud deal would be detailed next year. "As we look to the months ahead, we see two certainties: rapid change and compute shortages," he said.
Feldman said OpenAI's business with Cerebras grew from the second quarter. Cerebras still lacked the capacity to support a broad rollout of OpenAI's Ultrafast, a faster service tier previewed in August for its GPT-5.6 Sol model.
He added that data center availability remains a constraint across the industry. US policy changes, he said, could create greater challenges for development.
Partnerships
In July, Cerebras announced a partnership with AMD to pair AMD's Helios systems with its wafer-scale engine for inference, the stage where a trained AI model answers requests. Feldman said Cerebras' cloud now uses AMD graphics processors for prompt processing.
Last month, Oracle expanded its cloud AI chip deals to add Cerebras offerings. The hardware will run in Oracle data centers, with the inference service expected to launch in summer 2027, Feldman said.
