China’s National Bureau of Statistics reported that the country produced 279.8 billion integrated circuits in the first half of 2026, a 23.1% increase year-on-year. The figure translates to more than 1.5 billion chips manufactured every single day, a volume that places China’s semiconductor industry on a trajectory that was difficult to predict when the first rounds of US export controls took effect in 2022. The data is a useful corrective to the assumption, still common in Washington policy circles, that restricting China’s access to advanced chip-making equipment and foreign-designed semiconductors would slow the country’s overall chip production.
How Export Controls Accelerated Domestic Production Rather Than Constraining It
The production surge reflects several converging forces. Government investment in domestic semiconductor capacity has been sustained and substantial: the third phase of the National Integrated Circuit Industry Investment Fund, known as the Big Fund III, has deployed tens of billions of yuan into chip design, fabrication, and equipment companies since its establishment in 2024. Domestic demand for chips has grown rapidly as China’s AI industry has expanded.
The country’s AI sector now generates 1.2 trillion yuan in annual revenue and requires enormous quantities of memory, logic, and specialized AI chips to support inference and training workloads. And the export controls themselves have created a powerful incentive for Chinese technology companies to source domestically rather than rely on foreign suppliers whose availability is subject to US government decisions.
CXMT, DFSX, and the Emerging Domestic Chip Ecosystem
The most visible manifestation of this dynamic is CXMT’s trajectory. The company has grown from a minor player in global DRAM markets to the world’s fourth-largest producer with approximately 8% market share, and it is now raising $8.55 billion in an IPO that opened for investor subscriptions today. Its five-year, $2.94 billion supply agreement with Tencent, disclosed on the eve of the IPO, demonstrates that China’s largest technology companies are actively committing to domestic chip supply chains rather than treating them as a fallback option.
The pattern is likely to repeat: as CXMT’s production scales and its technology matures, more Chinese cloud and AI companies will find it rational to source from domestic suppliers, creating a self-reinforcing cycle of demand and investment.
The logic extends beyond memory chips. Huawei’s Ascend AI chip ecosystem has matured into a credible alternative to Nvidia’s H100 and H200 for domestic AI training and inference workloads, and the company debuted its Ascend 950 SuperPoD supernode system at WAIC 2026 this week. DFSX’s DF1000, announced last week, represents a fully domestic AI chip on a 14-nanometer process, not competitive with cutting-edge 3nm or 4nm designs, but functional for a range of inference tasks and entirely outside the reach of US export control enforcement. The accumulation of these developments suggests that China’s chip industry is building depth across multiple segments simultaneously rather than concentrating all resources on a single technological breakthrough.
The ASPI Critical Technology Tracker, which monitors China’s performance across 64 technology categories, found that China leads in 57 of them — including several semiconductor-adjacent areas such as advanced materials, quantum sensing, and high-performance computing. That breadth of technological leadership complicates the export control strategy, which was designed to target specific chokepoints in the semiconductor supply chain. When a country leads in the majority of critical technologies, the number of effective chokepoints shrinks.
The Limits of the Surge: Mature-Node Volume vs. Cutting-Edge Logic
None of this means that US export controls have been ineffective. China’s most advanced chip production still lags the global frontier by at least one to two process generations, and the absence of extreme ultraviolet lithography equipment from ASML continues to constrain the pace at which Chinese fabs can advance their process nodes. The H1 2026 production figure of 279.8 billion chips is dominated by mature-node devices, the kind used in consumer electronics, automotive systems, and industrial equipment, rather than the cutting-edge logic chips required for frontier AI training. The gap between volume and sophistication remains real.
But the trajectory is clear. China’s chip export revenues rose 111% year-on-year in the first half of 2026, reflecting both the volume growth and the improving quality of domestically produced chips. The country is building a semiconductor industry that is increasingly self-sufficient at the mature node level and increasingly competitive at the advanced node level. Export controls that were designed to freeze China’s chip capabilities in place have instead accelerated the investment and organizational focus required to overcome them.