How China’s Experimentation Approach Is Narrowing The AI Leadership Gap
AIThis post was created with the assistance of artificial intelligence (AI).

📊 Full opportunity report: How China’s Experimentation Approach Is Narrowing The AI Leadership Gap on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

TL;DR

China is making tangible progress in domestic chip manufacturing, especially in DUV lithography and multi-patterning. This experimentation approach is gradually closing the technological gap in AI hardware, but significant challenges remain before full commercial-scale production at advanced nodes.

China has begun mass-producing domestic immersion DUV lithography machines, capable of supporting 7-nanometer and potentially 5-nanometer chip nodes, marking a significant step in its semiconductor self-sufficiency efforts. This progress, confirmed by credible sources, indicates China is moving beyond prototypes toward scalable manufacturing, a development that could impact global AI hardware leadership.

Multiple credible accounts confirm that China is now producing domestic immersion DUV lithography systems tied to Huawei-linked firms and evaluated at SMIC, targeting advanced chip nodes. These systems are capable of multi-patterning to reach 7-nanometer and possibly 5-nanometer processes. Separately, a domestic EUV prototype has been reported, although it remains at the experimental stage. SMIC has demonstrated 7-nanometer production using older DUV tools with yields estimated around 20%, significantly below the 90% yields typical of leading global fabs. Huawei aims to produce over a million high-end AI-accelerator chips this year, reflecting China’s strategic push into AI hardware.

Despite these advances, experts note substantial hurdles remain. Yields at current domestic facilities are far below commercial viability, and the materials supply chain, especially for ultra-pure photoresist, remains heavily dependent on Japanese suppliers. Additionally, Chinese tools lag behind ASML’s technology by approximately four generations, with credible forecasts suggesting domestically made tools may not reach sub-10-nanometer production before 2030. The installed base of DUV tools also depends on Western maintenance services, creating a dependency that limits self-sufficiency.

At a glance
reportWhen: ongoing, with recent developments repor…
The developmentChina’s efforts in domestic lithography and chip fabrication are resulting in real, scalable manufacturing capabilities, signaling a shift in global AI hardware competitiveness.
AI DISPATCH · REALITY CHECK Forward-looking · 11 Aug 2026
China’s chipmaking, past the headlines
The Learning-by-Doing Wall

Every few weeks a headline says China cracked the last hard problem in chipmaking — and triggers alarm in one camp, triumph in the other. Both overreact, because both mistake a learning-by-doing problem for a copying problem. It isn’t one.

▲ Forward-looking · figures are point-in-time estimates
~20%
SMIC 5nm yield vs ~90% on EUV
~90%
Of high-end photoresist from Japan
4 gens
Domestic DUV lag behind ASML
~2030
Est. sub-10nm commercial, at earliest
01
Four walls behind the wall

“A machine exists” and “a machine makes advanced chips at scale, profitably, for years” are separated by a chasm — made of things that only accumulate with time.

Yield ~20% vs ~90%
The difference between a demo and a business. A process throwing away four of five dies is a science experiment. Closing it takes ten thousand small fixes, each learned by running wafers.
Materials ~90% JP
Even a perfect machine needs ultra-pure photoresist — the “film” of chipmaking — and China buys ~90% from Japan. You can build the camera and still can’t make the film.
Generational lag ~15 yrs
Domestic DUV lags ASML by ~4 generations — its tools of 15 years ago. Independent forecasts: no sub-10nm commercial production before ~2030.
Servicing 200+ tools
The installed DUV tools aren’t self-maintaining; multi-patterning drifts optics out of calibration. Servicing still runs through ASML. A borrowed capability, not an owned one.
02
A phase transition, not a footrace

In a race, a burst of speed closes the gap. In a phase transition, you can’t move faster to cross over — you have to accumulate enough, slowly, until the system changes state.

heat / capital / time in → state liquid — demos, prototypes the wall: tacit knowledge accumulates steam — commercial production
Water doesn’t become steam by heating faster. The capability arrives when the process has run long enough, at enough scale, fixing enough failures, that the unbuyable, untransferable know-how of how to actually do it has accumulated. ASML earned it over decades with TSMC, Samsung, Intel — China is building it largely in isolation.
03
How to read every headline

When you see “China achieves X,” ask which of two very different claims is actually being made.

Claim A
A machine functioned
A prototype made light. A tool made a few chips. A demonstration succeeded under controlled conditions.
vs
Claim B
Commercial production began
Sustained yield. Reliable uptime. Years of operation. An actual, profitable business at scale.
Almost all the real difficulty lives in the gap between A and B — and almost all coverage collapses them into one. The alarmist and the triumphalist make the same mistake.
04
The sober signals confirm the slow read

Even amid the loud headlines, the quiet data points all say the same thing.

Chinese media itself went quiet on tool progress and moved to deny an inflated 90% yield claim — insiders know the demo-to-production gap better than the headlines.
ASML’s China sales are falling as a share — yet China still can’t do without its tools, or its servicing.
The domestic machine ships in units of ~5 this year, ~20 next — real, and a rounding error against what one leading fab installs.
The gap is a wall, not a footrace — a phase transition of unbuyable know-how.
No prototype, no shipped tool, no yield headline teleports past it.

Implications for China’s Semiconductor and AI Hardware Leadership

This progress signifies a strategic shift for China in its quest for semiconductor independence, especially in AI hardware. Moving from prototypes to scalable, reliable manufacturing at advanced nodes could reduce reliance on Western technology and supply chains, potentially altering the global AI hardware landscape. However, the substantial technical and material challenges mean that China’s full capability at the most advanced nodes remains years away, and current progress is part of a long-term phase transition rather than an immediate leap.

The Hardware Hacker: Adventures in Making and Breaking Hardware

The Hardware Hacker: Adventures in Making and Breaking Hardware

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

China’s Semiconductor Ambitions and Technological Barriers

Over the past decade, China has prioritized developing its semiconductor industry amid export controls and geopolitical tensions. While it has achieved some success in manufacturing older and mid-range chips, the most advanced nodes—crucial for high-performance AI hardware—have remained out of reach due to technological, material, and supply chain barriers. The recent progress in domestic lithography tools and chip fabrication indicates a deliberate, sustained effort to overcome these barriers through experimentation and incremental learning, rather than rapid breakthroughs.

"China is making tangible progress in domestic chip manufacturing, especially in DUV lithography and multi-patterning, signaling a shift in global AI hardware competitiveness."

— Thorsten Meyer

The MACHINE that makes the MACHINES: Inside ASML and the race to control the future of microchips (AI)

The MACHINE that makes the MACHINES: Inside ASML and the race to control the future of microchips (AI)

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Remaining Challenges and Uncertainties in China’s Chip Capabilities

While China has demonstrated progress, it is not yet clear when domestic tools will reliably produce chips at sub-10-nanometer nodes at scale. The impact of material dependencies, especially for ultra-pure chemicals, and the ability to maintain high yields over time remain uncertain. Additionally, the extent to which China can develop independent maintenance and supply chains for advanced lithography equipment is still unresolved.

MANGTING Toner Chip Removal Tool – Chip Recycle Drill for use for hp 276/258/259

MANGTING Toner Chip Removal Tool – Chip Recycle Drill for use for hp 276/258/259

  • Model Compatibility: HP 276/258/259

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Next Steps in China’s Semiconductor Development Roadmap

China is expected to continue refining its domestic lithography and fabrication processes, aiming to improve yields and material independence. Industry sources anticipate that by 2025-2030, China could achieve commercial-scale production of 7-nanometer chips at higher yields, reducing reliance on Western equipment and materials. Monitoring government policies, technological breakthroughs, and supply chain developments will be crucial to assessing progress toward full self-sufficiency.

Amazon

domestic DUV lithography system

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

How close is China to manufacturing advanced chips at commercial scale?

While China has made progress in producing 7-nanometer chips at low yields and is developing 5-nanometer capabilities, it is still years away from achieving reliable, high-yield, commercial-scale production at the most advanced nodes.

What are the main technical hurdles China still faces?

Key challenges include improving manufacturing yields, reducing dependency on imported ultra-pure materials, and developing domestically capable lithography tools that match the performance of leading Western equipment.

How might this progress affect global AI hardware competition?

If China successfully scales its domestic chip manufacturing, it could significantly reduce reliance on Western supply chains, potentially shifting the balance of AI hardware leadership over the next decade.

Will China be able to produce chips at sub-10-nanometer nodes soon?

Most credible forecasts suggest that domestically produced tools capable of sub-10-nanometer production will not be available before around 2030, making this a long-term phase transition rather than an immediate shift.

Source: ThorstenMeyerAI.com

You May Also Like

Week Three — Foundation model vs Brownian motion. Kronos on five-minute BTC.

Testing of Kronos against a Brownian baseline on five-minute Bitcoin trades shows no significant advantage, challenging assumptions about modern models’ predictive power.

The SSD Squeeze: Why Storage Joined the Party

Enterprise and consumer SSD prices soar as AI drives unprecedented storage demand and wafer competition tightens supply, reshaping the market in 2026.

China: The Visible Hand

China’s government actively directs AI, robotics, and industrial growth through top-down planning, emphasizing state ownership and control over innovation.

Apple CEO confirms price hikes, Take Two announces GTA 6 preorder date

Apple CEO confirms upcoming price increases; Take Two announces GTA 6 preorder date, signaling significant shifts in tech and gaming markets.