Cerebras Q2 Revenue Doubles to $210M Amid Cloud Expansion

Cerebras (CBRS) posted a second-quarter 2026 loss of 4 cents per share, a result that exceeded the Zacks Consensus Estimate by 80.95 percent. Core revenues reached $209.87 million, doubling year over year and surpassing analyst expectations by 8.09 percent. The primary driver of this growth was the rapid expansion of its fast inference services, which shifted the revenue mix significantly toward cloud-based solutions.

According to data cited by GN markets/earnings (en-US), the stock has declined 17.2 percent since the earnings report was released, underperforming the broader S&P 500 index. While the top-line growth was substantial, the company continues to operate at a loss, with adjusted EBITDA at negative $53.1 million compared to negative $38.3 million in the prior-year period. The market’s reaction highlights the tension between accelerating revenue and the heavy capital expenditure required to support these cloud deployments.

Cloud Services Drive Revenue Growth

Core cloud and other services revenues surged 287 percent year over year to $127.7 million, nearly quadrupling from $33 million in the same period last year. This increase reflects the ramp-up of OpenAI’s deployment and higher usage from other cloud customers. In contrast, core hardware revenues grew a more modest 17 percent to $82.1 million, indicating a strategic shift in how the company monetizes its silicon and infrastructure.

Management noted that revenue mix can vary significantly quarter to quarter depending on the timing of large cloud-capacity additions and hardware shipments. Despite this variability, demand remains strong, with several late-stage hardware opportunities representing hundreds of millions of dollars in potential value. The company also reported that remaining performance obligations reached $25.4 billion, suggesting a substantial backlog of committed future revenue.

Strategic Partnerships Expand Inference Capabilities

Cerebras deepened its infrastructure partnerships by enabling support for OpenAI’s GPT-5.6 Sol at 750 tokens per second. The company also expanded its disaggregated inference strategy with AMD, pairing GPU-based prefill processing with Cerebras systems for decoding. Management stated that this configuration can maintain Cerebras’ speed while increasing throughput by up to fivefold, with production entry expected in the fourth quarter of 2026.

A similar disaggregated inference offering with AWS is expected to become generally available through Amazon Bedrock in the first quarter of 2027. The company also signed new cloud-capacity agreements with AI coding companies Cognition and Lovable. Additionally, six deals worth more than $30 million each were signed during the quarter, with customers including Block, Figma, AlphaSense, and GSK utilizing fast inference for agentic workflows.

Operating Margins and Balance Sheet

Core gross margin improved to 40.6 percent, up approximately 940 basis points year over year. However, sequential margin pressure was noted due to the temporary use of higher-cost systems rented back from cloud customers. Operating expenses totaled $502.8 million, with research and development accounting for $320.2 million. The core operating loss narrowed to $33.6 million from $43.9 million a year earlier, improving the core operating margin to negative 16 percent from negative 42 percent.

The balance sheet strengthened significantly, with cash, cash equivalents, restricted cash, and short-term investments totaling $8.6 billion as of June 30, 2026. The company has access to an $850 million revolving credit facility, which remained unused at quarter-end. Net cash used in operating activities for the first six months of 2026 was $47.5 million, a figure that management indicated supports ongoing capacity investments and operational scaling.