China’s EUV Lithography Breakthrough: Inside the “Manhattan Project” Defying US Chip Sanctions

For the last three years, the narrative in the semiconductor world has been one of “containment.” Washington builds a fence, the Netherlands locks the gate, and China is left on the outside, forced to “multi-pattern” its way into relevance using outdated DUV (Deep Ultraviolet) tools. CEOs, analysts, and the very engineers at ASML claim that building an Extreme Ultraviolet (EUV) machine from scratch was a twenty-year task requiring the collective genius of the Western world.

Well, it looks like someone forgot to tell Shenzhen.

Reports emerging this week from sources close to the project suggest that China has not just been “trying” to build an EUV machine; they have actually built a working prototype. Behind the high-security walls of a laboratory in Shenzhen, a massive, factory-floor-sized prototype is currently humming with 13.5nm light. If the photonic quantum chip we covered recently was a warning shot, this is a technological declaration of independence.

EUV vs. DUV: Why 13.5nm light is the game-changer

To understand the stakes, you have to understand the wavelength. Traditional DUV machines use 193nm light. It is a “thick Sharpie” trying to draw microscopic lines. To get to 7nm or 5nm chips with DUV, China has had to use Self-Aligned Quadruple Patterning (SAQP), essentially drawing the same line four times to make it thinner. It is slow, prone to errors, and results in yields as low as 33%.

EUV uses a 13.5nm wavelength, nearly X-ray territory. It allows for “single-exposure” printing of the world’s most advanced transistors. This light is so fragile that it is absorbed by air, meaning the entire process must happen in a total vacuum. Without EUV, you are not just slower; you are effectively locked out of the AI revolution.

How China recruited Ex-ASML engineers for a “Manhattan Project”

ASML’s EUV machines are often called the most complex devices ever built. They are the size of a school bus and cost $250 million. China’s version? It is reportedly much, much bigger, filling nearly an entire factory floor.

The most fascinating part of this leak is not the hardware; it is the tradecraft. Sources describe a “Manhattan Project” level of secrecy coordinated by Huawei and state research institutes:

• The “Ex-ASML” Workforce: The project is reportedly staffed by a specialized team that includes former ASML engineers recruited with signing bonuses up to $700,000, working under aliases to avoid detection by Western intelligence.
• The “Frankenstein” Feat: Unable to buy new parts, the team used components salvaged from older ASML machines purchased through secondary markets and middlemen, essentially re-engineering a working EUV light source from existing Western hardware.
• Total Sovereignty: The goal is clear: “China wants the United States 100% kicked out of its supply chains.”

From LPP to LDP: China’s alternative EUV path

While ASML uses Laser-Produced Plasma (LPP), hitting falling tin drops with a laser 50,000 times a second, China has pivoted to a “DeepSeek-style” optimization called Laser-Induced Discharge Plasma (LDP).

• ASML’s LPP: A miracle of timing. A high-power laser vaporizes a falling drop of tin in mid-air. It’s clean, efficient, but requires the ultra-high-power CO2 lasers and FPGA chips that the US has banned.
• China’s LDP: Vaporizes liquid tin between two rotating electrodes and uses a high-voltage electrical discharge (like a lightning bolt) to create the plasma.
  • The Advantage: It is mechanically simpler and uses a domestic supply chain.
  • The Trade-off: It generates massive heat and debris, which is why the prototype is so large; it requires a building-sized cooling and filtration system to protect the optics.

The Mirror problem: The last barrier to self-reliance

China has the light, but the “Final Boss” of EUV is the mirrors. In an EUV system, you can’t use lenses (they absorb the light), so you must use mirrors that are the flattest objects on Earth.

Zeiss mirrors are so smooth that if they were the size of Germany, the highest “mountain” would be 0.1mm tall. China currently lacks the “metrology” (the ability to measure these bumps) to match this. Instead, they are using “Chemical Smoothing” and Boron Carbide barriers to build mirrors that can at least survive the extreme heat of their LDP prototype.

The bottom line: 2028 or 2030?

The rest of the world has a refined, compact, and proven product. China has a massive, experimental laboratory. While ASML CEO Christophe Fouquet recently stated China would need “many, many years” to reach this point, the existence of a light-generating prototype suggests the timeline has shifted.

The Chinese government is reportedly targeting 2028 for the first functional chips. Independent analysts suggest 2030 is more realistic for high-yield mass production. Regardless, the “unhackable” monopoly of EUV has been cracked. Washington’s export controls were designed to buy the West time; if this Shenzhen prototype is real, that time just got cut in half.