China and US intensify AI rivalry over chips and models

The United States has begun treating advanced artificial‑intelligence models as matters of national security, warning Chinese firms of possible sanctions and accusing them of illicitly using U.S. technology. Officials allege that companies such as Moonshot AI employed “distillation” techniques to copy capabilities from models like Anthropic’s Fable 5, prompting investigations into illegal access to high‑end chips.

China is responding with a multi‑billion‑dollar push to reduce reliance on foreign semiconductors and accelerate domestic AI development. State‑backed initiatives have spurred the launch of new AI processors and open‑weight models such as Kimi K3, GLM‑5.2, and DeepSeek‑V4‑Pro. Chinese firms claim that performance gaps with U.S. counterparts are narrowing and that the country is on track to achieve self‑sufficiency in key AI technologies within a few years.

Analysts describe the competition as a new Cold War, with implications for global supply chains, Taiwan’s semiconductor industry, and Europe’s position in the AI hardware ecosystem. The dispute extends beyond chips to software stacks, model open‑source strategies, and the broader digital‑sovereignty agenda that will shape technology standards and security policies worldwide.