Live

Coming soon

Agriculture Tech
Autonomy
Avatars & Digital Humans
Biotech / Synthetic Biology
Blockchain / Crypto
Brain-Computer Interfaces
Climate Tech
Cloud & Edge Computing

Commerce
Cybersecurity
Defense
Digital Identity
Education Tech
Energy
Fashion & Textiles
Food Tech

Healthcare Systems
Longevity & Human Enhancement
Materials Science
Mobility
Quantum Computing
Smart Homes
Space Tech
Wearables

DeepSeek — China AI Labs · fobi

Live

Coming soon

Agriculture Tech
Autonomy
Avatars & Digital Humans
Biotech / Synthetic Biology
Blockchain / Crypto
Brain-Computer Interfaces
Climate Tech
Cloud & Edge Computing

Commerce
Cybersecurity
Defense
Digital Identity
Education Tech
Energy
Fashion & Textiles
Food Tech

Healthcare Systems
Longevity & Human Enhancement
Materials Science
Mobility
Quantum Computing
Smart Homes
Space Tech
Wearables

DeepSeek

DeepSeek is a Chinese AI research company founded in 2023 by Liang Wenfeng and spun out of his quantitative hedge fund, High-Flyer, which still owns and funds it. Based in Hangzhou, it develops open-weight models including the DeepSeek-V3 base model and the DeepSeek-R1 reasoning model, which drew global attention for matching frontier systems at far lower reported training cost. For its first two years the company turned down outside venture capital and was financed through High-Flyer; reports in 2026 indicated it was in talks for a first external funding round, but that round had not closed as of June 2026.

DeepSeek's open-weight V3 and R1 models showed a small, largely self-funded Chinese lab could approach frontier performance at a fraction of the reported cost, reshaping global expectations for AI training economics.

deepseek.com China AI Labs

Chinaprivate

At a glance

Total raised

Undisclosed

no public funding on record

Headcount

51-200

approximate · public range

Headquarters

Hangzhou

China

Founded

2023

3 years active

Status

Private

China AI Labs

Subsector

For subscribers

Subscribe to see the full profile

Including Funding rounds, Bull / Bear thesis, Stock + earnings, Roster changes, Patents, News, and Open roles.

Already subscribed? Sign in →

AI bull / bear

Bull case

DeepSeek demonstrated, with V3 in December 2024 and R1 in January 2025, that a 51-200-person lab could approach frontier performance at a fraction of the reported cost — the most consequential efficiency result of the modern AI era. Its structural advantages are unique: ownership and funding by the High-Flyer quant fund means patient, founder-controlled capital with no venture growth theater, and the discipline shows in a lab that declined outside money for two years while out-shipping vastly larger organizations. Open weights bought global developer mindshare no marketing budget could, making DeepSeek a default base model for a large share of the world's deployments and the reference point against which all model pricing is now negotiated. Reports in 2026 that it was in talks for a first external round imply it can now raise on terms of its choosing, if it chooses at all. Cost leadership is the one moat that compounds as a market commoditizes, and DeepSeek defines cost leadership in models.

Bear case

DeepSeek's strategy gives away its crown jewels: open weights mean every release is instantly absorbed by competitors, cloud providers monetize its models without paying it, and the efficiency techniques it publishes become everyone's cost curve within months. The commercial engine is correspondingly thin — a cheap metered API and a free app — with no disclosed revenue and no visible enterprise motion, while Western governments restrict its services and US export controls squeeze the compute pipeline its training cadence depends on. Dependence on High-Flyer concentrates risk in a single funding source tied to Chinese quant-market fortunes and a single decision-maker in Liang Wenfeng. The unclosed 2026 funding talks cut both ways: taking outside capital would dilute the independence that made the lab special, while refusing it leaves DeepSeek fighting rivals that raise more in a quarter than it has deployed in its lifetime. Being the world's favorite free model is influence, not a business.

Generated by editorial-sourced · June 12, 2026Refreshes quarterly · see disclosures