ASML sold “absolutely nothing” in Europe in 2026 — that is the phrasing in a Tom’s Hardware piece this week that had drawn 173 points and 484 comments on Hacker News by Sunday morning. An arresting number. It is being read as a verdict: on the EU Chips Act, on European industrial policy, on the continent’s ability to make anything that matters.
I’d like to submit that everyone reaching for the verdict has skipped the arithmetic.
Zero Is Not a Verdict
A net-zero regional sales figure is a bookings artifact before it is a demand signal. Leading-edge lithography tools are priced in the hundreds of millions, ordered years ahead, and delivered on schedules that slip a quarter without anyone’s strategy changing. Move one tool’s acceptance across a year boundary and a region’s headline number swings by more than most national tool budgets.
ASML’s own statement, as reported, asks Brussels to help “create demand.” That is the tell. A company that believed Europe had stopped buying wouldn’t be lobbying; it would be writing off the region and moving on. A firm asking for subsidies is a firm expecting orders.
None of which makes the number nothing. It makes it a question, and nobody has answered it: which orders slipped, which customers deferred, and whether the deferrals were capex timing or something worse. The headline has been through 484 comments of speculation and almost no reporting.
The Chokepoint Brussels Keeps Apologizing For
Here is what Europe actually owns, and it is not fabs. It is the equipment and materials layer — ASML in Veldhoven, Zeiss’s optics works in Oberkochen, and a supply chain of specialty chemical and metrology firms that no other region can assemble. Selling lithography systems to Taiwan, Korea, China, and the United States is the European semiconductor industry. It has been for twenty years.
The zero in the European column sits next to large numbers everywhere else. That isn’t decline. It’s a division of labor: Asia and America build the wafers, Europe builds the machine that builds the wafers. When your customers are offshore, your domestic order book goes quiet for reasons that have nothing to do with your competence.
There is a second irony worth naming. The moat around that machine is not a market outcome. It is the export-control regime the United States pushed and the Dutch and Japanese governments aligned with, restricting advanced tool shipments. The most protected industrial monopoly in the West is protected by states — which is a fact the free-market retelling of the ASML story tends to edit out.
Industrial Policy Works When It Has Exactly One Job
The EU Chips Act — roughly €43 billion, agreed in 2023 — set out to double Europe’s share of global chip production. That was the error, and it was an error of target selection, not of spending. It aimed at output share, a metric that measures how much concrete got poured, rather than at a chokepoint, a metric that measures whether anyone else can do the thing.
Compare the program that actually worked. EUV lithography exists because governments co-funded a consortium through years when no single firm would have carried the risk: one impossible machine, one lead supplier, a handful of customers, and a very long horizon. That is the rarest kind of industrial policy — narrow, patient, aimed at something nobody else could build. It should also make enthusiasts for unfettered markets flinch, since the crown jewel of European manufacturing is a state-shielded arrangement with excellent optics.
The EU’s answer to ASML’s number — help us create demand — inverts the problem. Demand for lithography machines does not come from a Commission program. It comes from chipmakers with order books fat enough to sign eight-figure capital commitments. You cannot subsidize a customer into existence.
The Reshoring Vanity Project
Which brings us to the part that ought to bother the right more than it does. The last five years of Western chip policy have been a competition to subsidize buildings: Intel’s Magdeburg project, the ESMC joint venture announced in Dresden in 2023 with TSMC, Bosch, Infineon, and NXP. Not one of those is a chokepoint. Each is a facility full of tooling a competitor elsewhere can also buy, staffed by engineers who can be hired away.
A procurement lead at a mid-sized Dresden tool supplier, standing in the lot after a supplier day, put the demand question more precisely than any white paper: “We’re not waiting on Brussels. We’re waiting on three customers to decide whether they’re building the second building.”
That is the whole argument in one sentence. Subsidies buy the first building. The second building is bought by customers who have customers.
So ASML sold nothing in Europe in 2026. Fine. The continent should stop treating the absence of fabs as a wound and start treating the presence of the machine as leverage — and it should stop writing checks for factories whose only distinguishing feature is a European postcode. Here is a test for the next round of chip money: does the subsidy buy a capability no other region could replicate within a decade? If the answer is no, you are not building an industry. You are building real estate.