Analysis
Founders Fund and Khosla Ventures have both sent partners on extended trips through China this month to meet AI founders in person, according to The Information, even as US venture investment into Chinese startups has fallen for years amid tightening capital controls on both sides.
I think this is the more honest story about US-China AI competition than most of what gets written about it. The policy conversation in Washington treats outbound capital restrictions and export controls as settled facts that simply route around China's AI industry. What these trips tell you is that the people whose job is to find the best returns on capital don't believe that's true anymore, and they're willing to spend weeks on the ground to check for themselves rather than take Washington's framing at face value.
The timing isn't neutral. Carnegie China's talent tracker, corroborated by this year's Stanford AI Index, found that the share of the world's top AI researchers working in China rose from 27.1% in 2022 to 40.6% in 2025 -- a genuine talent migration, not just a funding story. That landed the same week Chinese AI stocks fell on news that Beijing regulators are probing DeepSeek and Moonshot AI over allegations, made by Anthropic, that both companies routed customer requests through Claude to harvest its outputs. Both data points cut against a simple narrative: China's AI talent base is deep and growing even as its leading labs face fresh questions about how some of them got there.
“I think this is the more honest story about US-China AI competition than most of what gets written about it.”
What A Trip Like This Actually Buys A Fund
Pulse has previously covered Khosla Ventures across its AI portfolio bets. Neither firm has disclosed a new China-based investment resulting from these trips, and that's the point -- outbound investment rules restricting US capital flows into Chinese AI, semiconductor and quantum companies remain in force regardless of which partners fly to Shenzhen or Hangzhou. What a trip buys instead is direct information: what DeepSeek's, Moonshot's and Zhipu's actual technical roadmaps look like beyond what gets published, how far ahead or behind the frontier labs they really are, and whether the AI talent pool Carnegie just measured is producing genuinely novel research or fast-following US labs. For funds like Founders Fund and Khosla that write large checks into the same categories -- foundation models, AI infrastructure, robotics -- knowing what's actually happening in the world's other AI superpower is worth the trip even without a term sheet at the end of it.
Room for disagreement: it's entirely possible these trips are optics more than strategy -- a way for US funds to signal sophistication and hedge against being wrong about China's AI trajectory, without any intention or ability to actually deploy capital there given the regulatory environment. A partner meeting founders in Beijing generates a great LP update slide regardless of whether it changes a single portfolio decision, and neither firm has any disclosed mechanism for turning these relationships into investments even if they wanted to.
What I'd actually watch is whether either firm's portfolio companies start citing specific Chinese competitors by name in their own fundraising decks over the next two quarters. That's the tell that these trips changed how the funds underwrite competitive risk, not whether a press release about a China investment ever materializes.