DeepSeek vs Mistral AI: Revenue, Profit and Business Model
DeepSeek had $70.7M of annual revenue in FY2026. A full-year revenue figure is unavailable here for Mistral AI.
Latest financial snapshot
DeepSeek
- Latest annual revenue
- $70.7M (FY2026)
- Net income
- Unavailable here
- Net margin
- —
- Revenue growth
- +427.6% a year, FY2025–FY2026
Mistral AI
- Latest annual revenue
- Unavailable here
- Net income
- Unavailable here
- Net margin
- —
- Revenue growth
- —
Financial summary
DeepSeek
DeepSeek does not publish revenue or profit figures, and no audited numbers are available; earlier third-party revenue estimates could not be verified and have been removed from this profile. What is reported: the final DeepSeek-V3 training run cost about $5.6 million in rented-GPU terms (2.788 million H800 GPU hours, excluding research and hardware costs), and the company stayed funded by High-Flyer until June 2026. That month it closed a Series A of about 50 billion yuan (~$7.4 billion); Liang personally committed about 20 billion yuan, Tencent about $1.39 billion (10 billion yuan) and CATL about $695 million (5 billion yuan), according to Reuters. Post-money valuation was reported between ~$50 billion (Bloomberg) and ~$59 billion (SCMP). In July 2026 Bloomberg reported talks to raise ~$1.5 billion more at about $71 billion ahead of a mainland IPO targeted for 2027.
Mistral AI
Mistral is private and does not publish audited accounts. Reported figures: ARR of about $339 million (€300 million) in September 2025 (CNBC) and more than $400 million in January 2026, up from about $20 million a year earlier (CNBC Disruptor 50). CEO Arthur Mensch said in 2026 he expected ARR to exceed $1 billion by year-end. Funding history: ~$119 million (€105 million) seed (June 2023), ~$435 million (€385 million) Series A (December 2023), ~$678 million (€600 million) Series B at ~$6.55 billion (€5.8 billion) (June 2024), ~$1.92 billion (€1.7 billion) Series C led by ASML at ~$13.2 billion (€11.7 billion) (September 2025), $830 million of debt for data centres (March 2026) and a ~$3.39 billion (€3 billion) round led by Samsung at more than $23.7 billion (€21 billion) (September 2026). Mistral AI is Europe's premier artificial intelligence startup, rapidly achieving a private market valuation of roughly $6 billion backed by massive investments from Microsoft and a16z.
Revenue and profit by year
DeepSeek
| Year | Revenue | Net income | Margin | Growth | Source |
|---|---|---|---|---|---|
| FY2026 | $70.7M | — | — | +427.6% | Source |
| FY2025 | $13.4M | — | — | — | Source |
| FY2024 run-rate | $1.5M | — | — | — | Source |
Mistral AI
| Year | Revenue | Net income | Margin | Growth | Source |
|---|---|---|---|---|---|
| FY2026 run-rate | $400M | — | — | — | Source |
| FY2025 run-rate | ~$339M | — | — | — | Source |
| FY2024 run-rate | $16M | — | — | — | Source |
Where the revenue comes from
DeepSeek
No segment breakdown is published.
Mistral AI
- Enterprise API Token Consumption (La Plateforme)~45%
Usage-based per-million-token API fees paid by software developers and enterprises querying Mistral Large, Mistral Small, and Codestral.
- Cloud Hyperscaler Royalties (Azure, AWS, Snowflake)~35%
Revenue-sharing royalties collected when enterprise clients deploy and consume Mistral models natively inside public cloud marketplaces.
- On-Premise Private Sovereign Licenses~20%
Annual recurring multi-million-dollar software licenses for defense, government, and banking customers deploying models in air-gapped environments.
Business model and strategy
DeepSeek
How it makes money
DeepSeek makes money mainly by selling access to its models through a pay-per-token API (OpenAI-compatible endpoints such as deepseek-v4-pro and deepseek-v4-flash), priced well below US rivals. Its consumer chatbot on web and mobile is free. Model weights are published under the MIT license, so anyone can self-host them;
Growth strategy
DeepSeek's growth plan rests on three moves: keep releasing competitive open-weight models (V4-Pro, V4-Flash and the multimodal V4-Flash-Vision-Exp in 2026) to win developers; undercut closed-model API prices to grow paid usage; and reduce dependence on Nvidia by optimising for Chinese chips, shown by its September 2026 open-source toolkit for Huawei's Ascend accelerators.
Competitive advantage
DeepSeek's edge is cost efficiency: its models use sparse Mixture-of-Experts designs (V3 activates 37B of 671B parameters per token; V4-Pro reportedly 49B of 1.6T) and Multi-head Latent Attention, which let it train and serve near-frontier models with far fewer GPUs than US labs. Open MIT-licensed weights and low API prices give it unusually broad developer reach.
Mistral AI
How it makes money
Mistral makes money in four ways. It charges per token for API access to its models on La Plateforme; it sells enterprise licences and custom deployments, often on-premise or in private clouds, to companies and governments that want control over their data; it earns through cloud marketplaces such as Microsoft Foundry, AWS Bedrock and Google Vertex AI;
Growth strategy
Mistral's growth strategy combines sovereign infrastructure with enterprise partnerships. It is building data centres in France and Sweden, expanded its Microsoft partnership in July 2026 through a multibillion-dollar deal under which Microsoft uses part of its European compute, signed Accenture as a deployment partner in February 2026, and is moving into industrial AI through the May 2026 acquisition of Emmi AI and…
Competitive advantage
Mistral's main advantages are its position as the leading non-US, non-Chinese frontier AI lab, which suits European governments and regulated industries; open-weight models that customers can run on their own hardware; a research team led by founders who worked on Chinchilla at DeepMind and LLaMA at Meta;
Questions about DeepSeek vs Mistral AI
Which company has higher revenue — DeepSeek or Mistral AI SAS?
DeepSeek had $70.7M (FY2026). A comparable annual revenue figure for Mistral AI SAS was unavailable here.
How does revenue per employee compare for DeepSeek and Mistral AI SAS?
DeepSeek generates $442k / employee in revenue per employee. A comparable figure for Mistral AI SAS requires matching employee and revenue data from the same reporting period.
How do DeepSeek and Mistral AI SAS make money?
DeepSeek and Mistral AI SAS generate revenue in fundamentally different ways. DeepSeek: DeepSeek makes money mainly by selling access to its models through a pay-per-token API (OpenAI-compatible endpoints such as deepseek-v4-pro and deepseek-v4-flash), priced well below US rivals. Mistral AI SAS: Mistral makes money in four ways.
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 DeepSeek vs Mistral AI overview