Founders Fund and Khosla tour China’s AI labs — learning trips, not cheques - Dealroom

Despite a sharp contraction in U.S. venture capital flowing into China, top Silicon Valley investors are maintaining active on-the-ground reconnaissance of Chinese technology companies through extended learning trips. According to The Information, Founders Fund partners Sean Liu, John Luttig, and Joey Krug visited Chinese tech centers in the past year, including MiniMax's Shanghai headquarters and UBTech Robotics in Shenzhen. Khosla Ventures partners have conducted similar multi-year trips, with visits to Xiaomi in 2026 and Unitree in the prior year. Other notable U.S. venture firms including Dimension Capital, Costanoa, and Thrive Capital have also undertaken recent China diligence missions. Critically, these trips are framed as learning exercises rather than investment opportunities; the firms did not close new China-focused investments from these visits.

The regulatory backdrop explains the shift in VC strategy. U.S. venture investment into China-headquartered companies has declined dramatically to 105 deals year-to-date, down from 110 deals in 2025 and representing an approximately 80% contraction from 533 deals in 2021. This collapse reflects Biden-era outbound investment restrictions targeting semiconductors, AI, and quantum technologies, regulatory frameworks that remained in place through early 2025. These capital controls have effectively eliminated traditional venture check-writing as a viable strategy for U.S. VC exposure to Chinese innovation.

The focus of these learning trips reveals a shift in how Founders Fund and Khosla Ventures assess China risk. Rather than seeking equity upside, partners are evaluating portfolio supply-chain dependencies and direct competitor benchmarks across advanced manufacturing, robotics, autonomous vehicles, energy, and artificial intelligence. The trips target specific competitive concerns: China's cost advantage in open-source AI models (DeepSeek and Moonshot), ByteDance's venture capital arm (Seedance), and the concentration of humanoid robotics shipments among Chinese firms including Unitree and AgiBot, alongside China's dominance of key supply-chain components for humanoid systems.

The strategic implication cuts against public rhetoric. While prominent figures in the Thiel-adjacent Silicon Valley circles have voiced strong skepticism of China's technology sector, these named learning trips demonstrate continued institutional need to understand and monitor Chinese competitive capabilities and supply-chain concentration. The pattern suggests a durable bifurcation: public antipathy toward China investment coexists with private, methodical diligence of China's AI stack and manufacturing ecosystem. Regulatory restrictions on outbound capital have not eliminated the need for learning-based intelligence gathering; they have instead redirected it into in-person visits unencumbered by LP-facing China exposure or formal investment commitments.