As ASML reaches a ~€275B market cap, twice that of Intel, a look at the fundamental limits on making ever-smaller chips, which could finally end Moore's Law
For decades, ever-tinier semiconductors have allowed huge leaps in computing progress. Not for much longer, warn experts Bluesky: @revan.bsky.social . Tweets: @jasoncosta , @adam_tooze , @graphcoreai , @alexcampolo , @1br0wn , @newlefteviews , @pstasiatech , and @martijnrasser Bluesky: @revan.bsky.social : Another factor that people aren't considering often: electrical charge can only be discrete amounts. I think they bought some time when they moved to vertical 3D NAND, but the last generation of planar NAND were storing only a few electrons in the floating gate to change states. Tweets: Jason Costa / @jasoncosta : ASML story is amazing, and the Intel one has gotten sad. Hope the latter can turn it around. “.” https://www.ft.com/... Adam Tooze / @adam_tooze : “From its headquarters in Veldhoven, just a few kilometres away from that portacabin, ASML produces machines capable of vaporising tiny droplets of molten tin up to 50,000 times a second, creating a 13.5nm wavelength of light.” https://www.ft.com/... [image] @graphcoreai : What's the measure of progress for semiconductors post-Moore's Law? Graphcore's Nigel Toon and the boss of plucky upstart nvidia discuss the future in the FT (paywall). https://www.ft.com/... Alex Campolo / @alexcampolo : Spinoza ground lenses so ASML could vaporise... https://twitter.com/... Ian Brown / @1br0wn : “This idea of the singularity [when AI overtakes human intelligence] is bollocks, but the idea that you can use AI to create a next frontier of computing is perfectly practical.”-Nigel Toon, CEO @graphcoreai https://www.ft.com/... @newlefteviews : This is sort of why TSMC doesn't make Taiwan uniquely indispensable, since they will only be able to do EUV with ASML's lithography machines, which (and the article omits this) are only possible with Zeiss optics and Trumpf lasers. https://www.ft.com/... Paul Triolo / @pstasiatech : ASML has just five customers for its existing EUV machines — TSMC in Taiwan, Samsung and SK Hynix in South Korea, and Intel and Micron in the US. All of them have ordered the latest model... https://www.ft.com/... Martijn Rasser / @martijnrasser : ASML: “working on the first prototype of...newest product: the latest generation of EUV machines that will be used to “print” transistors almost as small as the diameter of a human chromosome on to sheets of silicon to make a semiconductor.” https://www.ft.com/...
Context & Ripple Effects
The valuation gap is a decade-long bet paying out. Back when Moore's Law first showed signs of slowing, leading chipmakers weighed adopting EUV lithography with mass production pencilled in for 2018; ASML was the only company willing to industrialize it, and the resulting chokepoint on cutting-edge chips is now priced into a ~€275B market cap, roughly double Intel's.
What makes this moment different from prior 'end of scaling' scares is that the warning now comes from physics rather than economics: last-generation planar NAND was storing only a few electrons per floating-gate cell, and vertical 3D stacking bought time rather than removing the limit. ASML's own answer — the High NA EUV transition it is preparing, with its €350M-per-unit high-NA machines already slated for TSMC, Intel, and Samsung and volume manufacturing expected around 2027–2028 — is both the extension of Moore's Law and evidence of how expensive each additional extension has become.
First-order effects
- ASML's five EUV customers — TSMC, Samsung, SK Hynix, Intel, and Micron — face steeply rising capital intensity per node just as demand growth becomes uncertain, a tension ASML's CEO has already flagged publicly over tariff-driven doubt about 2026 growth.
- Intel being worth half of ASML despite designing and manufacturing chips underscores that the scarce asset in semiconductors is no longer chip design or fabs ownership but the lithography toolchain itself.
Second-order effects
- Memory makers hitting discrete-charge limits push harder into 3D architectures, shifting competition from lithography resolution toward stacking and packaging — a different supplier base than the one ASML and Zeiss dominate.
- Each node extension concentrates more of the industry's cost structure into fewer suppliers — ASML, Zeiss optics, Trumpf lasers — making the entire leading-edge supply chain hostage to single-vendor execution and export-policy decisions.
Third-order effects
- If physical limits end transistor shrinage as reliably delivered gains, computing progress migrates from process geometry to architecture, specialized accelerators, and software efficiency — changing which companies capture value across the stack.
- A single-company chokepoint on the only path past those limits entrenches ASML's position as systemic infrastructure, drawing it permanently into geopolitics: whoever controls access to its tools effectively sets the frontier.
The trend: Moore's Law is transitioning from an economic rhythm sustained by shrinking transistors to a physical ceiling managed through exponentially costlier tooling, concentrating control of computing's future in one Dutch supplier.