At its September 2025 event, Huawei Connect, the company announced four new AI chip models to succeed the Ascend 910 series over the next three years . These chips underpin China’s domestic AI infrastructure, and Huawei is now considered the leader of China’s AI hardware stack—the foundational layer beneath the country’s booming large language model industry
. The roadmap is explicitly designed to work without US components, a strategic reality reinforced by government collaboration and deep partnerships with SMIC
.
Despite being cut off from leading US suppliers for years, Huawei reported $118.3 billion in revenue for 2024, a 22% year-over-year surge that marked its fastest growth since 2016 . The company now allocates 20.8% of its revenue to research and development, a rate that rivals the world’s most innovative firms
. It also successfully diversified into smart automotive solutions and expanded its global telecom equipment market share—moves that directly benefited from the self-sufficiency it was forced to build
.
The consequences go far beyond Huawei, revealing a systemic paradox in how Western export policy interacts with Chinese industrial strategy.
Export controls didn’t just block a few components—they severed access to electronic design automation (EDA) tools, advanced foundry services, and essential software, compelling Chinese firms to develop every layer of the stack locally . The Brookings Institution observed that this forced self-sufficiency was the exact opposite of the intended effect
. A peer-reviewed academic analysis concluded that America’s “chokepoint strategy is increasingly proving to be a fallacy,” arguing that US policies inadvertently accelerated China’s push for domestic innovation
.
Denied unlimited access to top-tier Nvidia chips, Chinese AI labs innovated around the constraint. They developed model architectures optimized for less powerful hardware, producing cheaper and more efficient AI systems that now rival Western models . By late 2025, Chinese models were regularly appearing in the global top ten across major benchmarks and were demonstrating a 90% cost-efficiency advantage over some Western counterparts
. The 3-to-6-month performance gap that once separated the two ecosystems now continues to narrow, turning a hardware weakness into a structural advantage
.
China’s semiconductor market reached $182.8 billion in 2024, with government targets aiming for 50% self-sufficiency in the near future . Massive subsidies, patriotic procurement policies, and coordinated national strategies have ensured that domestic alternatives like HarmonyOS and Ascend chips have a guaranteed market
. This isn’t just corporate survival—it’s an orchestrated national campaign to decouple from Western supply chains permanently.
The Information Technology and Innovation Foundation (ITIF) found that export controls actually helped Huawei while damaging US companies that lost access to Chinese revenue, and that the controls failed to halt China’s chip progress . While US policymakers tightened restrictions, Chinese firms accelerated their own production capabilities, effectively shrinking the market that American semiconductor companies can ever hope to recapture.
The technology gap hasn’t fully closed, and overstating China’s position would be a mistake. The US still retains a commanding lead in high-end AI hardware and the tooling that underpins semiconductor manufacturing. SMIC cannot import extreme ultraviolet (EUV) lithography machines, which are essential for producing chips at 3nm-class nodes and below . This equipment bottleneck remains a genuine barrier to closing the gap entirely
.
The sanctions bought the US time in the race for absolute cutting-edge performance, but at a clear cost: a fully independent Chinese tech ecosystem now exists, and it no longer depends on Western supply chains. That is the permanent structural shift Washington’s policy created.
Comments
0 comments