Mark RadarMARK RADAR
About
EN
Sign in

Cerebras Q2 Revenue Jumps 74% as AI Compute Services Surge

2 reports · First detected 2026-08-13 · Last active 2026-08-13

Cerebras Systems, a maker of wafer-scale AI processors that competes with Nvidia, is shifting from one-time chip sales toward cloud-based AI compute services. The transition is central to its effort to build more recurring revenue and improve operating leverage in a capital-intensive market, as demand for alternatives to dominant AI infrastructure suppliers continues to rise.

Cerebras reported second-quarter 2026 revenue of $180 million, up 74% from a year earlier, while cloud and AI compute-services revenue surged 281% and became the main growth driver. GAAP net loss widened to $451 million, largely reflecting stock-based compensation tied to its IPO, though its core operating loss narrowed after one-time items were excluded. The results missed market expectations, sending the shares down nearly 16% in after-hours trading.

All Coverage

2 original reports

The Backstory

The history behind this event
AI Chipmaker Cerebras Shares Fall 11% After First-Quarter Results2026-06-25 · 3 reports · similarity 0.81

Cerebras Systems competes in the AI training and inference market with its proprietary Wafer-Scale Engine, or WSE, directly challenging Nvidia’s GPU architecture. The company listed on Nasdaq on May 14, pricing its IPO at $185 a share and raising $6.4 billion. Its first earnings report as a public company became a key test of AI hardware demand, its lofty valuation and the cost of expanding capacity.

On June 23, Cerebras reported first-quarter 2026 results for the period ended March 31. Revenue rose 94% from a year earlier to $193.4 million, while its GAAP net loss was $14 million. However, the company forecast a second-quarter core gross margin of just 36%–38%, down from 47% in the first quarter. Its shares fell as much as 11% in after-hours trading, then dropped another 19.61% to close at $182.26 on June 24, below the IPO price.

Mark Radar|MARK RADAR
All times are in Taipei time (GMT+8)