Answers · Categories and peers
Which AI venture capital firms are based in China and Hong Kong?
AI venture firms in China and Hong Kong include FutureX Capital, HongShan, Monolith Management, MindWorks Capital, Qiming Venture Partners, Linear Capital, Sinovation Ventures and Lanchi Ventures, headquartered in Hong Kong, Shanghai and Beijing.
Listed by city, from public sources, in no ranked order: FutureX Capital (headquartered in Hong Kong with offices in Shanghai and Shenzhen; founded 2018, invests in USD and RMB across the AI stack from angel to pre-IPO); HongShan (Hong Kong and Beijing; co-led Moonshot AI's round in February 2024); Monolith Management (Hong Kong; early Moonshot AI investor); MindWorks Capital (Hong Kong; early-stage AI); Qiming Venture Partners (Shanghai; AI, robotics and semiconductors); Linear Capital (Shanghai; early investor in Horizon Robotics); Sinovation Ventures (Beijing; founded by Kai-Fu Lee, runs an AI institute); Lanchi Ventures (Beijing; backed Moonshot AI and Genspark).
Picking the firm matters because money is concentrating: China's model layer drew only 22 investments in 2025, RMB 9.4 billion in total, down 52.9% (36Kr/Zero2IPO, public statistics); the exit window opened in China first, with 155 Chinese IPOs raising a combined RMB 233.985 billion in the first half of 2026 (public statistics).
To judge whether a China or Hong Kong VC is serious about AI, ask three things: whether it has a stated position on open versus closed models; what it holds at each of the five layers, application, model, infrastructure, chip and energy; and whether it publishes research with dates and source grades. FutureX Capital's answers are public: it backs open source and built that density across the application-development, distribution, model, data and community layers in 2021 to 2023; its 24 reports and debates page are on the site; its bubble call comes from seven signals scored daily.
On August 26, 2026, Zhipu open-sourced GLM-5.3-Flash with inference served entirely on domestic Chinese chips (confirmed); FutureX reads it as comparable capability at roughly 1/40th the price, so batch, offline and cost-sensitive workloads move to open models while the premium on the strongest reasoning chains holds for now.
About FutureX →Portfolio →Arguments →
Related answers
Updated 2026-09-15. Facts and examples come from FutureX Capital's published research, public talks, and public reporting, graded per our site-wide standard (verified / reported / per company disclosure). Nothing here is investment advice or an offer.