Cerebras Q3: Cloud Now 63% of Revenue, Microsoft Signs On
Cloud services drove $144.9M of Cerebras's $230.9M Q3 core revenue, and Microsoft signed as a customer. Shares jumped 17% on a raised 2026 outlook.
Why it matters
- Cloud and services generated $144.9M of $230.9M Q3 2026 core revenue, about 63% of the total.
- CEO Andrew Feldman named Microsoft as the hyperscaler customer; revenue expected from Q1 2027.
- Cerebras raised its 2026 core revenue outlook to $940M–$960M; shares jumped ~17% after hours.
- The company holds $26.1B in remaining performance obligations and 700 MW of contracted data center capacity through 2027.
- GAAP net loss was $161.2M; core net income was $29.8M versus a $10.9M loss a year earlier.
Cerebras Systems now earns most of its revenue renting out its dinner-plate-sized chips, and Microsoft has become its newest cloud customer. Cloud and other services delivered $144.9 million of the company's $230.9 million in core revenue in the third quarter of 2026 — about 63% of the total — while core hardware revenue slipped to $86 million from $96.8 million a year earlier, the company said in its results release Thursday.
Chief Executive Andrew Feldman named Microsoft as the customer behind what Cerebras called "an important cloud deal with a top three hyperscaler." Microsoft signed in October, Feldman told CNBC. "We believe revenue begins to show up in Q1 '27," he said. Cerebras disclosed no value or terms for the Microsoft contract and said more details would come next year.
What does the Microsoft deal mean for Cerebras?
Microsoft joins Amazon Web Services and Oracle on the cloud customer list. Oracle plans to offer Cerebras inference inside Oracle Cloud Infrastructure in 2027. Landing a third hyperscaler validates Cerebras's shift from selling wafer-scale hardware to selling access to it — a recurring-revenue model that investors reward more readily than lumpy chip shipments.
The stakes are high for a company whose shares had fallen roughly 47% since its May listing on Nasdaq. The Microsoft announcement, combined with an earnings beat, reversed some of that slide: shares jumped about 17% in extended trading Thursday.
How fast is the business growing?
Core revenue rose 70% from a year earlier. That is slower than the 103% growth Cerebras posted in the second quarter, but it beat the $214.7 million analysts expected in LSEG data cited by CNBC. Core cloud revenue rose 272%. On a GAAP basis, cloud revenue grew 265%.
Cerebras now expects $940 million to $960 million in core revenue for full-year 2026, up from its earlier forecast of $880 million to $890 million, Reuters reported. For the fourth quarter, the company guided to $310 million to $330 million in core revenue with a core gross margin of 47% to 49% — both above analyst estimates of $271 million and 42.1%.
"We won new marquee customers, including hyperscalers, neoclouds and large enterprises," Chief Financial Officer Bob Komin said in the release. He added that the pipeline for high-speed AI inference grew more than tenfold during the quarter.
Is Cerebras profitable yet?
Profitability is closer but not achieved. The Q3 numbers break down as follows:
- Core gross margin: 41%
- Core operating margin: negative 14%
- GAAP net loss: $161.2 million, counting stock compensation and warrants issued to large customers
- Core net income: $29.8 million, against a loss of $10.9 million a year earlier
The gap between the GAAP loss and the positive core net income reflects the warrants Cerebras has issued to big customers — a structure that trades equity-linked dilution for committed demand. The swing from a $10.9 million core loss to $29.8 million in core income marks the clearest evidence yet that the services model can carry the company's cost base.
Where is the constraint: demand or supply?
Supply. Cerebras reported $26.1 billion in remaining performance obligations and said it has 700 megawatts of data center capacity under contract for delivery by the end of 2027, including new sites in Norway and Finland. The company announced the capacity figure alongside its Q3 highlights on X.
Reuters reported that the backlog is underpinned by Cerebras's deal to supply computing power to OpenAI, and that Feldman said on the analyst call Cerebras lacks the capacity to support a broad rollout of OpenAI's Ultrafast service tier. The company ended September with $7.8 billion in cash and investments plus $850 million in debt capacity. Feldman told CNBC the company is weighing financing options to keep up with demand.
The financing question now shapes the competitive field. Fellow AI cloud provider Nebius said Thursday that it asks customers for prepayments to cover build-out costs. Cerebras has not said it will follow that route, but its $7.8 billion war chest sits against a $26.1 billion backlog it must build data centers to fulfill.
What opening does Feldman see against Nvidia?
GPU supply itself, according to the CEO. "I think there is opportunity created by, particularly on the hardware side, the sale of hardware, where long lead times in GPUs and more expense creates opportunity for us," Feldman told CNBC.
The results landed the same day Nvidia named Colette Kress interim CEO after Jensen Huang resigned — a leadership shift at the dominant AI chip supplier that adds uncertainty to a market where long lead times already frustrate buyers. Cerebras positions its wafer-scale systems, sold as a service, as a way for hyperscalers and enterprises to sidestep GPU queues.
What comes next?
Cerebras expects to triple revenue in 2027. Its next-generation CS-5 system, built on the Wafer Scale Engine 5 processor now in system testing, is targeted for early summer 2027. Executives will present to investors Tuesday at the TomorrowX Summit in Austin. The first test of the raised outlook arrives with fourth-quarter results, when the guidance of $310 million to $330 million in core revenue must convert into booked growth — and when the market learns whether Microsoft revenue will begin to show up on the Q1 2027 timeline Feldman laid out.
Source: webpronews.com
More from Rebecca Stone
Show full bio
Correspondent covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
609 articles
Related articles
- Cerebras Raises Full-Year Outlook as Cloud Revenue Jumps 265%
- CoreWeave Revenue Jumps 168% to $3.661 Billion, Backlog Hits $132.5 Billion
- Microsoft Cuts Its Neocloud Roster as AI Buildout Shows Strain
- AWS Revenue Jumps 39% to $45.98 Billion, Fastest Growth Since 2021
- Microsoft’s 60,000-Job AI Buyout Drives Major Restructuring