Thirteen of China's top semiconductor executives published a joint target in March 2026: 80% domestic self-sufficiency in chips by 2030. The names on that list include the chairman of Yangtze Memory Technologies and the chairman of Naura Technology Group, two of the most consequential figures in Chinese semiconductor manufacturing. The signal is real. The gap between signal and reality is the story.

Before assessing whether the target is achievable, it helps to understand what the target is actually measuring, because "chip self-sufficiency" is not a single number. It spans design software, raw materials, manufacturing equipment, fabrication capacity, and packaging. China has made genuine progress in some of those areas. In others, the gap remains structural and is not closing quickly.

Why chips are hard to make

Every digital device runs on chips: tiny pieces of silicon covered in microscopic circuits. The smaller those circuits, the more powerful and efficient the chip. Making them requires a machine that prints those circuits onto silicon using light, much the way a printer transfers ink to paper. The difference is that the circuits are thousands of times thinner than a human hair, so the light source has to be extraordinarily precise.

The company that makes the only machines capable of printing the smallest, most advanced circuits is ASML, headquartered in Eindhoven, the Netherlands. There is no second source. Every chip inside an advanced smartphone, a cloud server, or an artificial intelligence accelerator was made on an ASML machine, or depends on components made on one.

ASML did not arrive at this position through luck. The company spent roughly three decades iterating through increasingly difficult physics problems to reach the extreme ultraviolet, or EUV, generation of its tools. The machines depend on a supply chain of more than 5,000 specialized subcontractors across dozens of countries who make the mirrors, lasers, sensors, and process control software. None of those components are interchangeable with off-the-shelf parts. They took years to develop to the tolerances required. You cannot reverse-engineer the machine without also reverse-engineering the supply chain and the accumulated manufacturing data that tells operators how to run it reliably at scale.

Where China actually stands

The US government and its allies have blocked ASML from selling its most advanced machines to Chinese customers. China has been working to build its own equivalent for years. The results, as described candidly by the same executives who signed the 80% target, are not there yet.

China's most advanced domestically produced deep ultraviolet lithography system is roughly equivalent to an ASML machine originally designed for 32-nanometer-class processes. Even if SMIC, China's leading chip foundry, integrates that tool into a 28-nanometer process by 2027, reaching sub-10-nanometer production, which is where the most powerful AI chips are made, requires a different category of machine entirely and several additional years of development beyond that.

A prototype EUV machine has reportedly been completed in a laboratory in Shenzhen. It was built by former ASML engineers using parts from older ASML machines purchased on secondary markets, and the Chinese government's target is to produce working chips on that prototype by 2028, with the development team itself reportedly viewing 2030 as more realistic. For context, ASML spent close to two decades solving the yield challenges required to make EUV commercially viable after its own prototype stage.

What China's plan actually contains

The 2030 plan is not solely about EUV. It has three more grounded near-term goals. First, a complete 7-nanometer production line equipped entirely with domestically produced tools by 2030. Second, stable volume production of 14-nanometer chips. Third, reducing dependence on foreign equipment across the mature-node segments, which cover chips used in automobiles, industrial controls, and consumer electronics, where China has made more meaningful progress.

The Chinese government already requires that any new fabrication plant installation contain at least 50% domestically produced equipment. SEMI, the industry association that tracks global semiconductor capacity, projects China's share of global production capacity to grow from 25% in 2024 to 42% by 2028. That projection is largely based on mature-node expansion, not leading-edge breakthroughs. The investment vehicle behind much of this, a state fund with roughly $47.5 billion earmarked for semiconductors, has redirected fresh capital specifically toward lithography and electronic design automation tools as substitutes for ASML and Synopsys products.

China's previous plan, Made in China 2025, set a 70% self-sufficiency target by 2025. The actual figure at the close of that period was somewhere between 14% and 33%, depending on whether you measure by unit volume or by value. The gap matters because cheap, high-volume chips inflate unit-count self-sufficiency while advanced chip dependency remains. Goldman Sachs estimates that even with current momentum, domestic suppliers will meet around 37% of China's semiconductor demand by value by 2030, well below the 80% headline figure.

What this means for enterprise procurement

For an executive building a multi-year infrastructure roadmap, the China chip story has a practical read that gets lost in geopolitical coverage.

China's mature-node production is real and growing. Chips used in networking hardware, power management, certain classes of industrial sensors, and embedded controllers are increasingly manufactured domestically at competitive cost. Any enterprise with significant hardware procurement, particularly in Asia-Pacific markets, will encounter Chinese-sourced components in that tier whether procurement policies acknowledge it or not.

The leading edge is a different situation. The AI accelerators, the advanced memory, and the high-performance processors that power cloud infrastructure remain dependent on fabrication processes that require ASML's EUV tools. China cannot yet manufacture those chips domestically. That gap is unlikely to close by 2030 at a scale that changes procurement options for enterprise AI infrastructure.

The relevant planning variable is not whether China hits 80% in aggregate. It is which segments of your supply chain sit on mature nodes versus leading-edge nodes, and how your hardware sourcing strategy accounts for a world where that bifurcation deepens over the next four years rather than resolves.