Chipmakers are already approaching ASML to reserve lithography systems for 2029 and 2030 output, the Dutch company's chief executive has confirmed. The disclosure lands while ASML is still working through how to build enough machines to satisfy the current AI-driven wave of orders.

The statement, reported by Tom's Hardware, frames an unusual problem for the semiconductor supply chain. The sole supplier of extreme ultraviolet lithography equipment says demand may exceed its capacity even though it has committed to a large expansion of its manufacturing footprint.

Why can the sole EUV supplier not keep up?

ASML occupies a singular position. No other company builds production-grade EUV scanners, and every advanced logic and memory fab depends on them. When AI datacenter spending accelerates, that pressure arrives at one supplier's order book first.

Chief Executive Christophe Fouquet said demand could surpass capacity despite the company's expansion plans. He did not attach a figure to the potential shortfall, and the 2029 and 2030 inquiries remain negotiations rather than binding bookings. The confirmed fact is the conversations themselves: customers want machine slots nearly half a decade out.

That lead time tells its own story. A tool ordered today for 2029 delivery reflects fabs that have already secured construction timelines, wafer supply and end-market contracts. The AI buildout is being planned in five-year increments, not quarters.

What does this mean for the equipment market?

The immediate consequence is pricing power. When a monopolist faces demand beyond its output, allocation becomes the governing mechanism. Customers with the largest committed volumes — the leading-edge foundries and memory makers serving AI accelerators — secure slots first. Smaller players wait.

The second consequence is investment discipline. Expansion capacity that ASML adds for this cycle must also serve any future correction. The company has been here before: the 2021-2022 shortage gave way to softness in mature-node demand, and tool orders fell. Fouquet's framing keeps that caution visible. He described risk, not a signed backlog.

Analyst projections for AI infrastructure spending run well into the hundreds of billions of dollars cumulatively through the end of the decade, though ASML did not endorse a specific number. The company's own guidance, and the customer interest it cites, are the firmer signals.

How solid are the 2029-2030 inquiries?

An inquiry is not an order. Chipmakers routinely explore delivery slots years ahead to preserve optionality, and some of these conversations will not convert. But the practice reveals how buyers now read the market: capacity, not capital, is the binding constraint.

The pattern also shifts leverage toward suppliers throughout the tool chain. If ASML is capacity-bound, the same logic applies to the component vendors, metrology firms and materials producers behind it. Fouquet's comment is, in effect, a forecast for the entire equipment tier.

The comparison is instructive. A fab planning 2029 output today is behaving like an airline buying aircraft slots before the manufacturer has confirmed the production line — confidence in future demand has outrun the physical ability to serve it.

What comes next?

ASML's expansion plans will determine whether the 2029 conversations turn into revenue or into a queue. The company has not said when its added capacity comes online, so the gap between demand and output remains unmeasured.

For customers, the practical question is allocation. For ASML, it is calibrating growth against a cycle that has already punished overbuild once. The next earnings report, with its booking figures, will show whether the 2030 interest is hardening into contracts.