China Accelerates A.I. Self-Sufficiency, Eroding U.S. Export Leverage

Researchers at the Center for Security and Emerging Technology (CSET) at Georgetown University reported that China is pushing aggressively to build an independent artificial intelligence supply chain, as private firms and state-linked companies develop domestic chip and model alternatives. The move is occurring against the backdrop of U.S. controls intended to limit China’s access to high-end chips, and CSET’s analysis says those controls are losing leverage as Chinese capacity grows.

The CSET piece highlights Chinese companies that are developing substitutes for U.S. hardware and foundational models, noting recent advances by firms that aim to replicate or replace capabilities previously supplied by American vendors. The think tank frames these developments as part of a coordinated industrial strategy that includes government support for domestic semiconductor design, localized AI model training, and incentives for firms to shift supply chains onshore.

Chinese industry moves to close key technology gaps

CSET’s reporting points to a mix of private startups and larger incumbents pursuing homegrown chip designs and AI models. These efforts span from specialized inference chips to full-stack model development, and they are driven in part by sanctions and export restrictions that have constrained access to the most advanced U.S. components. The result is a rapidly maturing domestic ecosystem that can sustain increasingly capable AI workloads without relying on the same foreign inputs.

U.S. export controls face diminishing returns

The think tank argues that U.S. export controls — designed to slow China’s progress by restricting sales of advanced processors, design tools, and AI training systems — are having a more limited effect as Chinese alternatives appear.

CSET’s analysis suggests that sustained investment, supply-chain rearrangement, and targeted industrial policy inside China are allowing some companies to substitute away from blocked technologies, which reduces the leverage Washington expected to gain from controls.

CSET researcher Jacob Feldgoise is cited in the piece for explaining how these shifts undercut a core U.S. strategy: using export limitations to shape the pace and direction of Chinese AI development. The analysis traces how incremental engineering wins and domestic procurement policies can blunt export-driven pressure and lengthen the timeline for policy impact.

Policy implications for Washington

For policymakers and national security officials in Washington, the CSET assessment signals a need to reassess assumptions about how export controls translate into strategic advantage. The think tank calls attention to the limits of unilateral restrictions and the possibility that long-term competitive dynamics will rely more on alliances, investment in domestic resilience, and adaptive regulatory tools than on export denial alone.

Beyond immediate trade measures, the analysis points toward broader questions about where to invest public resources, how to coordinate with allied governments, and which technologies will remain chokepoints in the medium term.

What to watch next: observers should track follow-on technical demonstrations from Chinese firms, shifts in Beijing’s procurement and subsidy programs, and any changes in U.S. export policy or allied coordination that aim to restore leverage.