律动BlockBeats|8月 31, 2026 04:15
**[Latest Judgment from U.S. VC After Visiting China: AI Capabilities Need to Cross the Pacific Twice to Be Sold to U.S. Clients]**
Beating AI Newsflash: U.S. venture capital firm Dimension spent a week in China conducting research, visiting AI labs, investors, and entrepreneurs. Upon returning, they wrote a 5-page internal letter to their LPs. Dimension had already conducted research in Shanghai last year, but this time they visited both Beijing and Shanghai to recalibrate their assessment of Chinese AI.
Dimension highlighted that Chinese open-weight models have already entered the production chain of U.S. AI companies. Cursor's Composer 2 continued training from Kimi K2.5, and legal AI company Harvey's Tenet underwent further training on Kimi K3. They summarized this process as "crossing the Pacific twice": U.S. labs first train cutting-edge models, Chinese labs absorb these capabilities and release open-weight models, U.S. application companies then use the Chinese models for further training, and finally, products are created and sold to U.S. enterprises.
However, while usage has increased, the revenue has not proportionally stayed with Chinese model companies. Dimension summarized it in one sentence: "China is getting the West's workload, but not the West's revenue." Open-weight models can be deployed on different platforms, with U.S. inference companies like Fireworks, Baseten, and Modal taking the lion's share of service fees. Chinese models drive costs down, ultimately benefiting U.S. cloud providers, inference platforms, and AI application companies.
Dimension's final judgment is that it has become increasingly difficult to separate U.S. and Chinese AI into two entirely independent systems. While hardware like chips is decoupling—U.S. constrained by electricity and grid issues, and China lacking advanced chips—models, data, inference services, and software frameworks continue to flow across borders. Today, simply debating "who is winning in U.S.-China AI" no longer accurately reflects the reality of the industry chain.
[Original Article Link]
Share To
Timeline
HotFlash
APP
X
Telegram
CopyLink