Who's who · Chinese labs

DeepSeek

Hangzhou AI company known for low-cost, MIT-licensed open-weight models such as DeepSeek-R1 and V4.

At a glance

Founded
2023
Based in
Hangzhou, China
People
Liang Wenfeng, Founder
Known for
DeepSeek V4, DeepSeek-R1, DeepSeek-V3, DeepSeek app
Licence
MIT
Funding
Private; valued at more than 350bn yuan, about $52bn (July 2026)
First external round
About 50bn yuan, July 2026
Website
https://www.deepseek.com

DeepSeek is a Chinese AI company that develops large language models and publishes most of them as open weights under the permissive MIT licence. It was founded in 2023 by Liang Wenfeng, who also co-founded the quantitative hedge fund High-Flyer, and for its first years it was funded without outside investors.

DeepSeek became widely known in January 2025, when its R1 reasoning model matched leading US models at a fraction of the reported training cost and briefly wiped hundreds of billions of dollars off NVIDIA's market value. Its current family is DeepSeek V4, released as a preview in April 2026. V4 Pro has 1.6 trillion parameters and V4 Flash 284 billion; both are mixture-of-experts models, which use only part of their parameters for each task, and both have context windows of one million tokens.

In July 2026 DeepSeek completed its first external funding round, raising about 50 billion yuan at a valuation of more than 350 billion yuan (about $52 billion), with Tencent and CATL among the investors and Liang Wenfeng putting in money himself. In February 2026 Anthropic accused DeepSeek of using Claude to train its own models, a practice known as distillation.

Sources

  1. DeepSeek (Hugging Face), DeepSeek-V4-Pro, 2026-04-24
  2. TechCrunch, DeepSeek previews new AI model that 'closes the gap' with frontier models, 2026-04-24
  3. Caixin Global, DeepSeek Reaches $52 Billion Valuation in Round Backed by Tencent, CATL, 2026-07-17
  4. TechCrunch, Anthropic accuses Chinese AI labs of mining Claude as US debates AI chip exports, 2026-02-23

Facts last checked 2026-10-02