Cerebras Systems Inc.
Key statistics
from XBRL data in SEC filingsAI briefing
from the latest 10-K, 10-Q and 8-K eventsCerebras Systems Inc. is a wafer-scale AI processor and cloud inference company that recently went public and is scaling capacity under large contracts with OpenAI and AWS.
What they do
Cerebras designs and commercializes the world's first and only wafer-scale processor, delivering AI compute for training and inference. It sells AI supercomputers for on-premises deployment and offers compute via Cerebras Cloud and partner clouds (AWS Marketplace, Microsoft Marketplace, IBM watsonx, Vercel, OpenRouter, Hugging Face). The company also provides AI services to co-develop solutions and maintain on-prem hardware.
Revenue drivers
- Cloud and other services revenue — GAAP cloud and other services revenue was $126.0M in Q2 2026, up 281% YoY; core (non-GAAP) cloud revenue was $127.7M, up 287% YoY, and is the fastest-growing segment.
- Core total revenue (non-GAAP) — Core total revenue was $209.9M in Q2 2026, up 103% YoY, excluding non-cash warrant amortization, stock-based compensation, and data center pass-through costs.
- OpenAI master relationship agreement — OpenAI committed to purchase 750MW of inference compute capacity, with deployment through 2028; revenue recognition began in Q1 2026 and the initial warrant tranche vested upon a $1.0B working capital loan.
- Data center capacity under contract — Over 600 MW of data center capacity is live or under contract for delivery by end of 2027, with a pipeline in gigawatts, supporting future revenue growth.
Recent performance
In Q2 2026, GAAP total revenue was $180.1M, up 74% YoY, while core total revenue reached $209.9M, up 103% YoY. GAAP gross margin was 14% and core gross margin was 41%, up ~940 bps from Q2'25; core operating margin improved ~2600 bps to (16%). The company ended Q2 with cash and equivalents of $6.74B (per balance sheet) and total assets of $11.63B. Remaining performance obligations stood at $25.4B.
Strategy
Cerebras is scaling data center capacity to over 600 MW under contract and plans to more than triple revenue in 2027. It is expanding manufacturing via new factory lines at Flex, Sanmina, and Rocket EMS to increase capacity more than 10x in 2026. The company is deepening partnerships with AWS and AMD for disaggregated inference, aiming to bring 5x throughput benefits to Amazon Bedrock in Q1 2027. It is also growing customer base in AI coding (Cognition, Lovable) and security (CrowdStrike) to expand fast inference adoption.
Risks
- Customer concentration and dependency on OpenAI — A substantial portion of revenue and remaining performance obligations is tied to the OpenAI MRA; failure to deliver or collect could materially impact financials.
- Early-stage commercial execution — The company has a limited operating history and is scaling production and data center buildout rapidly, which may strain operations and supply chain.
- GAAP losses and margin pressure — GAAP operating margin was (265%) in Q2 2026 due to non-cash items; core operating margin is also negative, indicating significant costs in advance of revenue.
- Dependence on third-party manufacturing and supply — Reliance on TSMC for wafers and on contract manufacturers (Flex, Sanmina, Rocket EMS) for assembly creates supply chain risks if capacity or quality issues arise.
Outlook
For Q3 2026, management guides core revenue of $214–216 million, core gross margin of 38–40%, and core operating margin of (25%) to (23%). For full-year 2026, the company expects core revenue to grow significantly, planning to more than triple revenue in 2027. Management plans to use the $850 million revolving credit facility and IPO proceeds to accelerate data center acquisitions.