Cerebras improves full-year outlook with cloud revenue growth of 265% – Unite.AI

Cerebras Systems raised its full-year revenue forecast on November 12, 2026, after third-quarter GAAP revenue increased 55% year-over-year to a record $210.8 million. Revenue from cloud and other services rose 265% to $141.9 million, growth the company attributed to demand for high-speed AI inference, but its GAAP net loss widened to $161.2 million.

Cerebras is now targeting $940 million to $960 million in full-year core revenue, rising from $880 million to $890 million in its second-quarter release on Aug. 12, 2026. Its fourth-quarter outlook calls for a core operating margin of between negative 1% and positive 1%, bringing the adjusted measure close to breakeven. The key results exclude stock-based compensation and data center pass-through items and add recorded non-cash customer collateral amortization versus revenue, among other adjustments.

Cloud services drive revenue mix

Cloud and other services provided the majority of the quarter’s GAAP revenue, while hardware revenue declined from a year earlier. This shifts much of Cerebras’ growth toward providing inference capabilities as a service, integrating direct system sales and third-party cloud access.

Cerebras reported $26.1 billion in performance obligations remaining as of September 30, 2026, up from $25.4 billion as of June 30, 2026. The scale of these commitments makes deployment schedules and available capacity critical to turning customer demand into reported revenue.

Cerebras also announced that Oracle is integrating its fast inference into Oracle Cloud Infrastructure and that it has signed a cloud agreement with an unnamed customer described as one of the three major hyperscalers. Oracle is expected to be available in 2027, and more information on the hyperscaler deal is expected that year.

GAAP losses and infrastructure spending

GAAP net loss compared to $22.2 million a year earlier. Stock-based compensation accounted for $155.1 million of the adjustments in the third-quarter operating reconciliation, helping to explain why Cerebras’ core operating result is much closer to breakeven than its GAAP result.

Expansion also requires a lot of liquidity. During the first nine months of 2026, Cerebras used $388.9 million in operating assets and purchased $1.046 billion in property and equipment. It ended in September with $7.8 billion in cash, cash equivalents, limited liquidity and short-term investments, providing substantial resources to finance construction.

Capacity provision becomes the next test

Cerebras has secured 700 gross megawatts of data center capacity to be delivered by the end of 2027, including new contracts in Norway and Finland. The August release separately described more than 600 active or contracted megawatts to be delivered by the same deadline, a figure that also included capacity already operational.

Production is expanding in parallel with Flex and Sanmina. Cerebras expects to emerge from 2026 with more than 10 times the production capacity it had at the start of the year, with plans for additional factories and capacity in 2027.

The company says its disaggregated inference architecture can increase the number of tokens served from a data center by up to five times, improving the potential revenue produced per megawatt. This capability is critical to its effort to provide high-speed inference without requiring a commensurate increase in powered infrastructure.

The improved outlook links Cerebras’ next phase to execution as much as demand. Converting customer commitments into distributed inference capacity, while managing cash usage and GAAP losses, will determine whether rapid cloud growth produces sustainable computing economies.