China ramps up efforts to build domestic AI chips in response to U.S. restrictions, with state funding and tech giants leading the charge.
State-Driven Push for AI Chip Independence
China has initiated a major program, often called the “Manhattan Project” for AI chips, to gain semiconductor self-sufficiency.
This effort responds to tightened US export controls on advanced chips, updated through December 2025, which limit China’s access to key technologies.
On December 8, 2025, the U.S. eased rules to allow Nvidia H200 shipments to China, but broader restrictions persist.
Plans for up to $70 billion more in incentives were reported on December 12, 2025.
These funds support the full supply chain under the Made in China 2025 plan, which aims for 70% self-sufficiency in semiconductors by 2025, though progress lags.
Blending Emulation and Domestic Advances
China combines chip reverse-engineering with original development. In early 2025, a Shenzhen team completed an EUV machine prototype.
Efforts include high-security labs where engineers, some from Western firms, dissect foreign equipment from October 24, 2025.
To build expertise, China recruits global talent for semiconductors, offering high salaries and visas.
The K-visa, launched October 1, 2025, targets top STEM professionals.
Huawei plays a central role in coordinating AI chip development, expanding its Ascend ecosystem by November 21, 2025.
Startups like DeepSeek develop AI models using domestic chips, with its V3.1 release on August 21, 2025, optimized for Huawei hardware.
Persistent Hurdles in the Global Race
China faces ongoing challenges, including dependency on foreign tools and materials from April 21, 2025.
A talent shortage of up to 200,000 workers persists, per industry estimates. Geopolitical tensions add pressure.
This strategy narrows the gap with the West, reshaping tech dynamics in 2025.
