Cerebras Systems Inc. (CBRS)

NVIDIA LPX beat Cerebras’s deployed CS-3-era endpoint—not necessarily its contemporaneous CS-4 platform. LPX and CS-4 are arriving in essentially the same product window, so the real race remains unresolved. Measured Gemma 4 31B speed: LPX delivered 3,382 tok/s/user at 10k context, versus approximately 1,402 tok/s for Cerebras’s public endpoint—a...

Cerebras Gemma 4 31B is a latency product, not a budget inference product. At $0.99/M input + $1.49/M output, it costs about 6× other Gemma hosts but generates roughly 10–27× faster. Same-model alternatives: DeepInfra, Friendli, and SiliconFlow cost about $0.0021/request; Cerebras adds roughly 1.08¢ per request ($10.82/1k; $10,820/1M). What the premium...

Cerebras Big Chip Club interview with Ro Varma, OpenAI product lead for Codex, enterprise, and the Cerebras work. Background: engineer/founder, Palantir, Cursor; joined OpenAI in 2026. Quotes below are from that conversation, lightly cleaned of filler ("um," repeated "like") but not paraphrased. This is a product interview, not a contract or capacity...

As of August 20, 2026. Ordered from most important to least important using a subjective blend of disclosed relationship size, apparent compute commitment, and the size/strategic importance of the customer organization. This is not a revenue ranking. “Production use” means the organization publicly uses Cerebras technology, but the available source does...

- OpenAI is simultaneously Cerebras’s anchor customer, construction lender, strategic partner and potential low-teens shareholder. - Committed capacity: - 250MW by year-end 2026 - 500MW cumulative by year-end 2027 - 750MW cumulative by year-end 2028 - OpenAI may option another 1.25GW by 2030, reaching 2GW total. - Agreement value: more than $20B. Hardware...

- 2026 OpenAI commitment: 250 MW by year-end. - At 23 kW per CS-3: ~10,900 CS-3-equivalent systems/WSEs. - At 1.2–1.3 facility PUE: ~8,400–9,100 systems before other IT equipment. - End-2027 company capacity: more than 600 MW live or under contract. - Mechanical maximum: ~26,100 CS-3-equivalent systems/WSEs. - At 1.2–1.3 PUE: ~20,100–21,700 systems...

Cerebras has demonstrated that wafer-scale processors can be manufactured and deployed commercially, and OpenAI's GPT-5.6 Sol Ultrafast preview is a meaningful validation of the architecture for frontier-model inference. The strongest evidence supports a narrower conclusion than the broadest promotional claims: Cerebras has a genuine low-latency decode...