OPINION: Light in Shenzhen, Darkness in Europe

This week Reuters reported that China has assembled a prototype extreme-ultraviolet lithography machine inside a high-security Shenzhen lab—an apparatus that can generate EUV light, sprawls across a factory floor, and was reportedly assembled by reverse-engineering Western know-how and recruiting former ASML engineers.

The headline temptation is to declare export controls dead. The adult conclusion is harsher: this is not, yet, a technological defeat. It is a counterintelligence defeat.

Call it “reverse-engineering.” Better: call it theft.

A machine that produces EUV light is not a machine that produces chips; between the two sits the world of industrial modernity—precision manufacturing, reliability engineering, metrology, yield, throughput—where physics becomes process, and process becomes power. If China can only make light, then controls have bought time. If China can make light reliably at production scale, then controls have begun to lose their coercive bite. The distinction is everything.

Delay is a strategy.

The Reuters detail that matters most is the one people least like to say out loud: the vector is not smuggled crates; it is human beings. “Components” can be embargoed. Knowledge walks.

Which brings us to the Manhattan Project comparison that Reuters’ sources invoked—useful, but only if we read it correctly. In 1942, the United States chose to treat an abstract scientific possibility as a concrete strategic emergency, created an organization with authority, secrecy, and money, and delivered a working weapon in 1945. The lesson is not that democracies can out-mobilize autocracies; it is that institutions win when they convert fear into method.

Light is not a weapon.

China’s EUV effort—if the reporting holds—resembles Manhattan in mobilization and differs in outcome: it has produced a prototype that emits the right wavelength, not a tool that can earn its keep in a fab. That “gap” is not a footnote; it is the whole story. ASML’s long slog from early prototypes to commercial chips was not a triumph of theory. It was a triumph of intolerant engineering: systems that must repeat perfection thousands of times, under thermal stress, in vacuum, at insane power density, without drifting a hair.

But if you are Europe, the more immediate parallel is not Los Alamos; it is Veldhoven—small town, giant monopoly, strategic chokepoint—and the baffling choice to treat it like a normal corporate success story instead of what it is: critical infrastructure in the age of algorithmic war.

Paper warnings are not walls.

Here the Dutch—and, by extension, the European Union—have been practicing a sort of administrative casuistry: write the threat assessment, publish the report, attend the conference, and assume the problem has been “managed.” Yet the very Reuters narrative includes the indictment: intelligence warnings issued after key engineers had already left, after recruitment pipelines had matured, after patents had been filed, after the center of gravity had shifted. A security service that can describe a fire but cannot stop the arsonist is not a service; it is an archive.

Dutch intelligence has said publicly what polite Europeans prefer not to hear: Beijing’s most reliable technology pipeline runs through people—researchers, engineers, mid-career specialists—quietly recruited out of sensitive sectors. Read that and the rest should stop sounding like “compliance” and start sounding like security: in strategic industries, a hire and a resignation are not mere HR calendar entries but movement on the map; privacy rules that bind only one side become a gift to the side that does not reciprocate; and court victories that cannot be enforced abroad are pageantry—useful for headlines, useless for deterrence.

Deterrence is not a press release.

Europe now murmurs about “screening” and “policies” as if the verbs themselves were a defense. The fix is more concrete and more morally honest: treat certain technologies as strategic assets, build a professional counterintelligence capability around them, and accept a narrow but real tradeoff—less institutional naiveté in exchange for preserving the very openness Europeans claim to cherish. Liberty does not require innocence. It requires boundaries.

Meanwhile, the commercial wing of Europe performs its own slow-motion sabotage. European—and especially German—automakers have perfected the art of flattering Beijing while calling it “market access,” as if a state that cartelizes law, finance, and information were merely a demanding customer. The irony is almost classical: the same continent that lectures the world on “values” has spent years training, employing, and credentialing technical talent—then watching that talent convert into a subsidy for its principal strategic rival.

Markets do not deter.

Here is the scenario policymakers should force themselves to picture—not a sci-fi leap to Chinese parity, but a grimly plausible sufficiency. 2030. China’s domestic EUV tool is still inferior, yields are still ugly, and costs are still extravagant. Yet it works well enough to produce advanced chips for missiles, drones, signals intelligence, and sanctions-proof communications. Export controls remain “successful” on paper—no ASML machines in China—while their strategic leverage has quietly decayed in practice. What, then, was the West buying with time it refused to use?

History is an early warning system.

The proper response is not panic; it is posture. Tighten secondary-market enforcement, yes—but do not pretend that policing auction houses is the decisive battle. The decisive battle is talent, process discipline, and speed. The West must assume that whatever it builds today will be studied tomorrow; its edge lies in widening the lead, not romanticizing the moat. That requires R&D at scale, industrial capacity at home, and a security mindset that treats the departure lounge as seriously as the loading dock.

If Reuters’ story is true, the turning point is not that China has conquered EUV. The turning point is that Europe has discovered—again—that openness without defense is not virtue; it is vulnerability.

Unchecked provocation invites repetition.