Cerebras vs CoreWeave: Revenue, Profit and Business Model
Cerebras reported $510M of revenue in FY2025 and $87.9M of net income. CoreWeave reported $5.1B of revenue in FY2025 and a net loss of $1.2B.
Latest financial snapshot
Financial summary
Cerebras
Cerebras reported revenue of $510 million in 2025, up about 76% from roughly $290 million in 2024, and swung to net income of $87.9 million from a $485 million loss. After its May 2026 IPO it began reporting 'core revenue', which excludes pass-through revenue and amortization of customer warrant assets. Q1 2026 core revenue was $191.3 million (up 92%), and Q2 2026 core revenue was $209.9 million (up 103%), against GAAP revenue of $180.1 million. Q2 GAAP net loss was $450.5 million, mostly from $386.6 million of stock-based compensation tied to the listing. Management raised 2026 core revenue guidance to $880-890 million and guided Q3 to $214-216 million. Remaining performance obligations stood at $25.4 billion.
CoreWeave
CoreWeave's revenue was $229 million in 2023, $1.92 billion in 2024 and $5.13 billion in 2025, with a 2025 net loss of about $1.17 billion. Q2 2026 revenue was $2.575 billion versus $1.212 billion a year earlier; adjusted EBITDA was $1.51 billion (59% margin) but GAAP net loss widened to $626 million because of $640 million of net interest expense. The company raised 2026 revenue guidance to $12.4-$13.2 billion and expects an annualized revenue run-rate of $18.5-$19.5 billion by year-end. Total debt was roughly $35.1 billion at June 30, 2026, and market capitalization was about $44 billion in mid-September 2026.
Revenue and profit by year
Cerebras
| Year | Revenue | Net income | Margin | Growth | Source |
|---|---|---|---|---|---|
| FY2025 | $510M | $87.9M | 17.2% | +75.7% | Source |
| FY2024 | $290.3M | -$485M | -167.1% | +268.9% | Source |
| FY2023 | $78.7M | — | 0.0% | — | Source |
CoreWeave
| Year | Revenue | Net income | Margin | Growth | Source |
|---|---|---|---|---|---|
| FY2025 | $5.1B | -$1.2B | -22.7% | +167.9% | Source |
| FY2024 | $1.9B | -$863M | -45.1% | +736.2% | Source |
| FY2023 | $229M | -$594M | -259.4% | — | Source |
Where the revenue comes from
Cerebras
- Cloud & Other Services (Fast Inference)
~61% of Q2 2026 core revenue
Per-token API fees and dedicated capacity contracts for running AI models on Cerebras systems in leased data centers, led by OpenAI. $127.7 million in Q2 2026.
- Hardware Systems
~39% of Q2 2026 core revenue
Sales of CS-3 systems and clusters to sovereign AI programs, universities, national labs, and enterprises. $82.1 million in Q2 2026.
CoreWeave
- Committed GPU capacity contracts
Majority of revenue
Multi-year take-or-pay contracts for dedicated GPU clusters with AI labs, hyperscalers and enterprises.
- On-demand compute
Not disclosed
Usage-based access to GPU and CPU instances for shorter training, fine-tuning and inference jobs.
- Storage, networking, inference and software
Not disclosed
AI object storage, managed inference (over $100 million ARR as of Q2 2026) and developer tools such as Weights & Biases.
Business model and strategy
Cerebras
How it makes money
Cerebras makes money in two ways. First, it sells hardware: CS-3 systems built around the WSE-3 chip, bought by sovereign AI programs, national labs, universities, and enterprises that want to own their compute. Second, and now larger, it sells cloud and services: customers pay to run models on Cerebras systems hosted in data centers it leases, either through per-token inference APIs or dedicated capacity contracts.
Growth strategy
Cerebras' growth strategy centers on fast inference. It is building out leased data center capacity to serve OpenAI and other model providers, distributing its service through Amazon Web Services under a March 2026 agreement, and pricing low-latency token generation at a premium to GPU clouds.
Competitive advantage
Cerebras' edge is its wafer-scale architecture. A GPU cluster splits a model across many chips and spends time and energy moving data between them over external links. The WSE-3 keeps 900,000 cores and 44 GB of SRAM on one piece of silicon, so weights and activations move across the wafer at 21 PB/s.
CoreWeave
How it makes money
CoreWeave makes money by renting GPU compute, plus storage, networking and software, to AI developers. Most revenue comes from multi-year committed contracts in which a customer reserves a dedicated cluster and pays for it whether or not it is fully used; a smaller share is on-demand usage.
Growth strategy
CoreWeave grows by signing large multi-year contracts first and then financing and building the capacity to serve them. In 2025-2026 it expanded deals with OpenAI (initially about $11.9 billion over five years, later enlarged), Meta (including a $21 billion expansion announced in April 2026) and Anthropic, and broadened its enterprise base with customers such as Caterpillar, Bentley Systems and Databricks.
Competitive advantage
CoreWeave's edge is speed of deployment at scale. It is consistently among the first clouds to bring up new NVIDIA systems (first commercial GB200 NVL72 in early 2025, first GB300 NVL72 in July 2025, and first Vera Rubin NVL72 bring-up in 2026), and its Kubernetes-native platform and Mission Control software are built only for GPU clusters. NVIDIA is also a shareholder, adding a $2 billion investment in January 2026.
Questions about Cerebras vs CoreWeave
Which company has higher revenue — Cerebras Systems Inc. or CoreWeave, Inc.?
Cerebras Systems Inc. reported $510.0M (FY2025), while CoreWeave, Inc. reported $5.1B (FY2025). By last reported revenue, CoreWeave, Inc. is the larger business, with Cerebras Systems Inc. reporting a smaller revenue base.
What is the market cap of Cerebras Systems Inc. vs CoreWeave, Inc.?
Cerebras Systems Inc.'s market capitalisation stands at $46.3B, while CoreWeave, Inc.'s is $44.0B. Cerebras Systems Inc. carries the higher market valuation, reflecting investors' expectations of its future earnings power relative to CoreWeave, Inc..
Which is more financially efficient — Cerebras Systems Inc. or CoreWeave, Inc.?
Cerebras Systems Inc. generates $720k / employee in revenue per employee, while CoreWeave, Inc. generates $2.34M / employee. CoreWeave, Inc. shows higher revenue efficiency per headcount, though this reflects business model differences — capital-light software companies routinely outperform labour-intensive manufacturers on this metric.
How do Cerebras Systems Inc. and CoreWeave, Inc. make money?
Cerebras Systems Inc. and CoreWeave, Inc. generate revenue in fundamentally different ways. Cerebras Systems Inc.: Cerebras makes money in two ways. CoreWeave, Inc.: CoreWeave makes money by renting GPU compute, plus storage, networking and software, to AI developers.
Which company is valued higher relative to revenue — Cerebras Systems Inc. or CoreWeave, Inc.?
On a price-to-sales (P/S) basis, Cerebras Systems Inc. trades at 90.8x P/S and CoreWeave, Inc. at 8.6x P/S. Cerebras Systems Inc. commands a higher revenue multiple, typically indicating that investors expect faster growth or higher future margins compared to CoreWeave, Inc.. A higher multiple is not inherently better — it may also signal that the stock is priced for perfection.
Is Cerebras Systems Inc. bigger than CoreWeave, Inc.?
By last reported revenue, CoreWeave, Inc. ($5.1B (FY2025)) is the larger company compared to Cerebras Systems Inc. ($510.0M (FY2025)). Revenue scale is one dimension of size — market capitalisation, employee count, and geographic reach are also relevant depending on the context.
Figures come from each company's filings and the sources linked beside them. Amounts reported in another currency are shown in US dollars at an approximate rate and marked with ~. Back to the Cerebras vs CoreWeave overview