DeepSeek

DeepSeek is a Chinese artificial intelligence laboratory developing open-weight large language models, founded in 2023 by Liang Wenfeng and spun out of the quantitative hedge fund High-Flyer, which Liang also co-founded. Liang controls close to 90% of the company, which until 2026 had grown entirely without outside venture capital — funded by High-Flyer's balance sheet and its substantial GPU compute cluster.

DeepSeek became globally significant in January 2025 when its DeepSeek-R1 reasoning model matched frontier Western models on maths, coding, and reasoning benchmarks while being trained at a small fraction of the reported cost, triggering a major repricing of AI infrastructure expectations across global markets. In April 2026 the lab released its V4 generation as open-source: V4-Pro, a 1.6-trillion-parameter Mixture-of-Experts (MoE) system, and V4-Flash, a smaller 284-billion-parameter variant optimised for cost-efficient inference.

DeepSeek's defining characteristic is engineering efficiency — architectural innovations such as multi-head latent attention and aggressive MoE sparsity allow it to deliver frontier capability at dramatically lower training and inference cost than dense models of comparable quality. All flagship models are released under permissive open-weight licences, enabling enterprises and researchers to self-host, fine-tune, and audit the models without sending data to a third-party API.

The company distributes models through its own chat application, an OpenAI-compatible API, and major model hubs. Founder Liang Wenfeng has stated the company's goal is artificial general intelligence (AGI) while keeping models open.

In 2026 DeepSeek entered talks for its first external funding round — reported at up to roughly $7 billion and a valuation in the $45–50 billion range, with Chinese state-backed investment vehicles including the National AI Industry Investment Fund and 'Big Fund III' reported to be involved — marking a strategic shift toward scaling compute and commercial revenue. DeepSeek is widely regarded as the leading open-weight LLM lab outside the United States and the most influential Chinese foundation model developer.