DeepSeek is a Chinese AI research lab founded in July 2023 by Liang Wenfeng, who is also CEO of quantitative hedge fund High-Flyer. Headquartered in Hangzhou, the lab is owned and primarily funded by High-Flyer and has built a reputation for releasing open-weights frontier models at training costs far below those of leading Western labs. Its V3 and R1 releases in late 2024 and early 2025 triggered a sharp re-evaluation of global AI compute and capex assumptions.
In April 2026, DeepSeek released a preview of its V4 series, including DeepSeek-V4-Pro at roughly 1.6 trillion parameters and DeepSeek-V4-Flash at around 284 billion parameters. Both models support a one-million-token context window and are distributed under the MIT license, allowing commercial use and self-hosting. The release reinforced DeepSeek's position as the most influential open-weights lab outside the United States.
After years of internal funding from High-Flyer, DeepSeek opened its first external fundraising round in 2026. Multiple reports describe the company as in talks to raise up to roughly four billion dollars at valuations that have moved from around ten billion dollars to as high as fifty billion dollars within weeks, with China's state-backed Big Fund and Tencent named as potential lead and parallel investors. As of mid-2026 these rounds had not closed and remained in discussion.
DeepSeek's models are available through its own chat interface and API at deepseek.com and via many third-party providers and inference hosts, with API pricing that is consistently among the lowest for frontier-class models. The lab is most often used by developers, researchers, and cost-sensitive teams who want strong reasoning and coding capability with open weights and predictable inference economics, rather than buyers who need enterprise-grade SLAs or data-residency guarantees outside China.