DeepSeek AI
Open-Source Frontier LLMs
- Valuation
- Not publicly disclosed
- Total Funding
- ~$7.4B(RMB ~50B) in its first outside round, Jun 2026, at a $50B+ post-money valuation; previously backed solely by High-Flyer Capital
About DeepSeek AI
DeepSeek AI is a Chinese AI research lab founded by quantitative hedge fund High-Flyer Capital that shocked the global AI industry in January 2025 by releasing open-source models (DeepSeek R1, V3) that match GPT-4-class performance at a fraction of the training cost. Their work demonstrated that frontier AI is achievable with far less compute than previously assumed.
Company History
Founded in 2023 by Liang Wenfeng, founder of High-Flyer Capital, China's largest quant hedge fund. The lab had access to a large Nvidia H800 cluster acquired before US export controls. DeepSeek V3 (Dec 2024) was trained for ~$5.6M — orders of magnitude cheaper than US frontier models. DeepSeek R1 (Jan 2025) matched OpenAI o1, went globally viral, caused a market panic, and triggered US Congressional AI competitiveness hearings. Their efficiency techniques (MoE, multi-head latent attention, FP8 training) have influenced the entire industry.
Company Timeline
6 milestones- 2023Founded by Liang Wenfeng as offshoot of High-Flyer Capital hedge fund
- May 2024DeepSeek-V2 demonstrated MoE efficiency at shocking cost
- Dec 26, 2024DeepSeek-V3 released — trained on 2.8M H800 GPU-hours for ~$5.6M
- Jan 20, 2025DeepSeek-R1 released — matched OpenAI o1 on benchmarks; app hit #1 globally
- Jan 27, 2025US market reaction: Nvidia stock fell 17% in a single day; Congressional hearings called
- Mar 2026DeepSeek V3-0324 update; DeepSeek R2 reported in development
Products & Services
4- DeepSeek V3
- DeepSeek R1
- DeepSeek Coder V2
- DeepSeek Chat App
People & ecosystem
Key People
- Liang Wenfeng (CEO/Founder, also High-Flyer Capital founder)
Key Investors
- High-Flyer Capital