Theses›Semiconductor manufacturing›High-NA EUV

High-NA EUV

High-NA EUV lithography reaches widespread HVM availability by 2026, enabling sub-2nm logic nodes; manufactures $200B+ chips annually by 2030 as <10nm reaches 20-30% of total semiconductor production.

  • Bull
  • Shut
  • $32B by 2035

The callHigh-NA volume comes 2028-2030, not 2026, and every layer of its supply chain is owned; the only startup prize is a challenger way to print the finest layers.

D1D2D3R1R2R3C1foragainst D1D2D3R1R2R3C1foragainst
Fig. 1The balance. Drivers and catalysts against restraints, each block sized by its weight; catalysts hang dashed because they have not happened yet, sized by their odds. It tips toward the drivers.
  1. D1Below 2nm every AI chip eventually needs High-NA patterning; standard EUV can only follow by stacking exposures, which costs yield and time at every node.
  2. D2Intel is past 1,000,000 wafers on High-NA and takes 14A to risk production on it, pulling scanner, resist and inspection spend forward.
  3. D3Sovereignty money wants a second source for the finest layers, which could fund X-ray or atom-beam challengers.
  4. R1Every layer is owned or gone: ASML the scanners, Zeiss the optics, Lasertec inspection; Irresistible Materials took TOK money, one Finnish materials company is public, Nearfield raised €135M.
  5. R2TSMC won't use High-NA in volume through 2029, and until cost parity around 2030 it is a penalty most fabs push back.
  6. R3A challenger tool would need a qualified critical layer at a leading fab, which nobody has done, after years of qualification.
  7. C1Intel's 14A goes into risk production on High-NA, with TSMC's A14 following in 2028 if it changes its mind.2027
Rests onthe beliefs underneath
Supported by · confidence
Threatened by · confidence
What would prove itdated tests, open and resolved
20272028202920302031today234567predictionswatch 20272028202920302031today234567predictionswatch
Fig. 2Each ring is a dated test: dashed while open, struck through when overdue, ticked or crossed once resolved.
  1. 1High-NA EUV reaches widespread availability by 2026.Dec 2026 · resolved wrong
  2. 2By end-2027 Intel runs High-NA EUV on >=1 critical layer in 14A risk production AND TSMC publicly commits High-NA insertion for A14 or earlier; either part missing keeps the thesis contested.Dec 2027 · open
  3. 3Intel running at least one critical 14A layer on High-NA, and TSMC publicly committing insertion for A14 or earlier.Dec 2027 · open
  4. 4Sovereignty capital funding X-ray or atom-beam challengers at the $40-100M scale.Dec 2028 · open
  5. 5High-NA EUV manufactures $200B chips annually by 2030.Dec 2030 · open
  6. 6<10nm semiconductor production reaches 20-30% of total volume by 2030.Dec 2030 · open
  7. 7Chips made on High-NA above $200B a year, and sub-10nm at 20-30% of wafer volume.Dec 2030 · open
Expressionslisted, private, technologies
Private · 2 private companies.

Irresistible Materials · xLight

3 more private companies tracked, not shown

Technologies
Who has already bet on itprivate rounds, from public sources

$40M raised by 1 company, $420M of it in the last three years. Wave building: Money is arriving: recent rounds and tier-1 funds are in, but no runaway leader yet.

Backed by a top-tier venture firm: xLight.

  • $40MxLight
    • Founded 2021
     Series B, Jul 2025 $40M (Playground Global, Boardman Bay Capital Management, Morpheus Ventures)
What changedthe record, newest first
  • 9 Jun 2026My 2023 call that High-NA EUV would be in widespread high-volume production by 2026 resolves wrong. One fab has a tool installed, the biggest is skipping it for its next node, and three fabs cannot ramp to volume in 6 months. The lesson: when the newest tool costs $400m and the cheaper one still works, fabs run the old one far longer than the roadmap slides admit.
Reading
Show the reading