In October 2022, the Biden administration launched what officials described as a “new approach” to protecting American technological supremacy. Export controls targeting advanced semiconductors and the equipment used to make them were meant to freeze China out of the frontier, buying time for US firms to consolidate their lead. Three and a half years later, the results tell a different story. As of January 2026, China’s domestic semiconductor equipment self-sufficiency had surged to 35%, up from roughly 25% just two years ago — a pace that would have seemed implausible to most analysts when the controls were first introduced.
The controls did cause real disruption. They pushed up costs, fractured supply chains, and delayed China’s access to the most advanced fabrication nodes. Nobody serious disputes that. But disruption and defeat are not the same thing, and Beijing has treated every tightening as a procurement order from the state.
The Acceleration Nobody Planned For
SMIC’s key technological achievement was developing a 7nm process using older deep ultraviolet lithography, validated by the 2023 launch of Huawei’s Kirin 9000S chip. From 2025 to 2026, the strategy shifted to scaling advanced AI chip production, including mass production of Huawei’s Ascend 910B and a planned doubling of 7nm capacity. That is not the behavior of an industry being strangled. It is the behavior of one that has adapted.
Chinese chipmakers, led by SMIC and its principal customer Huawei, are acquiring older-generation deep ultraviolet immersion lithography tools and repurposing them to fabricate near-frontier chips. The process is inefficient and the cost curve is punishing. But access to those older machines is allowing Chinese chipmakers to scale production far faster than US policymakers had anticipated.
The yields remain imperfect. The 7nm process at SMIC suffers from reported yield rates of between 20% and 40%, indicating the technology is not yet globally competitive and relies on a protected domestic market and state subsidies to be viable. But “not yet globally competitive” is very different from “contained.” The notion of 70% semiconductor self-sufficiency by 2028 is attainable in some but not all areas, and all of this is occurring on a timetable absolutely inconceivable to anyone working in the Chinese semiconductor industry before October 2022.
The financial architecture explains why. SMIC’s aggressive expansion is almost entirely underwritten by a state-directed investment strategy. It is a primary beneficiary of China’s national semiconductor funds, including the $47.5 billion “Big Fund III” launched in May 2024, which finances its advanced node development to overcome US sanctions. American chipmakers do not have that backstop. When export controls cut into their revenues, the losses are real and compound over time.
The Revenue Problem Washington Keeps Ignoring
The semiconductor industry saw 19% growth in 2024 with sales reaching $627 billion, and projections show the market will expand to $697 billion in 2026. China is one of the largest buyers in that market. When American firms are barred from selling there, the revenue that would have funded next-generation R&D simply disappears. Beijing’s state-backed firms do not face that constraint. The asymmetry is structural, not incidental.
Washington has started to acknowledge this. In December 2025, NVIDIA was permitted to resume shipping H200-class chips to China under a case-by-case licensing review, reversing the blanket denial that had been in place. National security hardliners objected. But the reversal signals a grudging recognition that blunt restrictions on commercially available hardware carry limited security benefit while imposing real costs on the American industry they are meant to protect.
The controls were never as precisely calibrated as their architects claimed. Hardware with legitimate civilian applications was bundled with restrictions on genuinely military-relevant technology. The result was a broad instrument sold as precision policy.
What the CHIPS Act Can and Cannot Fix
The domestic investment side of the strategy has produced more concrete results. The CHIPS Act brought enormous investment into the US from top semiconductor firms around the world, spurring an estimated over $110 billion in 2019 dollars — more than the entire amount of real investment in electronics and computers manufacturing from 2007 until 2020. In March 2025, TSMC announced it would invest $100 billion more in its Arizona site, including additional fabs, packaging facilities, and a research center.
That is real money building real capacity. But the CHIPS Act cannot substitute for the revenue base being eroded by overbroad commercial restrictions. You cannot simultaneously cut the funding streams of American chipmakers and promise to out-innovate a state-directed competitor with an essentially unlimited runway.
The rise of China’s semiconductor equipment sector is a clear indicator of a broader trend: the permanent bifurcation of the global technology landscape. What started as a series of trade disputes has evolved into two distinct technological stacks. China’s progress in self-reliance suggests that the era of a unified, globalized semiconductor supply chain is ending.
The Distinction That Policy Has Failed to Make
The semiconductor competition is real and worth taking seriously. CSIS analysts broadly agree that the United States retains genuine leads in frontier chip design, advanced manufacturing equipment, and the research ecosystem that generates next-generation capabilities. The question worth asking is whether current policy protects that lead or gradually undermines it.
A more effective framework would distinguish between transactions that genuinely threaten security and those that simply involve commercial competition. A Chinese civil engineering firm buying legacy chips for infrastructure is not the same risk as the People’s Liberation Army acquiring GPU clusters for weapons simulation. Policy that treats them as identical is not rigorous. It is reflexive. And reflexive policy, aimed at a state-directed adversary with a decades-long time horizon, is the kind that tends to lose.
Original analysis inspired by Imran Khalid from Foreign Policy In Focus. Additional research and verification conducted through multiple sources.
