Chinese AI startup Moonshot seeks more Nvidia chips amid compute shortage

China's AI sector is navigating a critical bottleneck. U.S. export controls have restricted access to the most advanced Nvidia chips needed for cutting-edge AI model training, creating a compute shortage across Chinese tech. Yet rather than simply accepting these constraints, companies like Moonshot are demonstrating resourcefulness by pooling capacity from multiple cloud providers—a practical workaround that keeps development moving. Concurrently, China's domestic semiconductor push is yielding measurable results: the 2,850% first-half profit jump for Chinese chipmakers signals that local production is ramping up and becoming economically viable. The shift becomes even more significant with reports of state-backed lithography machine production. Lithography tools are the capital-intensive equipment needed to manufacture chips at scale; if China can produce these domestically, it reduces reliance on suppliers like ASML and strengthens the entire domestic supply chain. The sharp sell-off in ASML shares, along with contagion selling in Japanese and South Korean chip stocks, shows that global investors now see this as a structural shift—not a temporary shortage but a potential reordering of who controls chip supply.