ASML-TSMC 12-Inch Photomask Shift and Its Supply Chain

On September 8, 2026, ASML and TSMC announced a joint industry initiative to move semiconductor photomasks from today's 6x6-inch format to a 6x12-inch large format, targeting a pilot line in 2031 and lithography readiness for advanced-node production in 2033. The 12-inch photomask shift could force a generational replacement across the entire mask supply chain. [1][2]

The two companies formed the collaboration ahead of the SPIE BACUS photomask conference, describing the goal as maximizing the value of High NA EUV lithography. TSMC said it intends to use ASML's High NA systems in high-volume manufacturing for advanced nodes starting in 2030. [1] Separately, Samsung and TSMC have committed to High NA EUV in high-volume manufacturing by 2028 and 2030 respectively, joining Intel as the third such user. [3] ASML closed at $1,764.85 that session, up 2.91% from September 4; TSMC's ADR closed at $439.00, up 2.35%. [4]

What a photomask is, and why the format matters

A photomask, or reticle, is a quartz glass plate carrying the pattern for one circuit layer of a chip. The lithography tool projects light off that plate and shrinks the image onto the wafer. An advanced chip has dozens of layers, and each needs one or more masks.

Today's reticle is a 6x6-inch square, and the entire supporting ecosystem was engineered around that dimension: blanks, mask writers, etch tools, inspection equipment, pellicles, and the sealed pods and in-fab transport systems that move masks around. [5]

High NA EUV is the next generation of extreme-ultraviolet lithography. Its resolution is higher, but a single exposure covers only half the field of the prior generation. A full chip design therefore has to be split into two half-fields, exposed separately and stitched together, which takes two masks per layer. Enlarging the mask is what removes the stitching.

The new format is a 6x12-inch rectangle, not a scaled-up square. Mask area roughly doubles, but the aspect ratio changes, so fixtures, chucks, carriers and optical stages are all new geometry.

A specification change, not a volume change

The chain starts with a specification change rather than a demand change. Industry coverage of the announcement states that because the whole mask ecosystem is built around 6x6, an entire new set of blanks, tools and technologies optimized for 6x12 has to be developed — and no party has said who pays for that R&D. [5]

The first landing point is capital spending. Existing 6x6 mask lines and inspection tools stop being the leading-edge asset from 2033, which means a second round of investment falls on suppliers whose current spend is 6x6.

The second is mask consumption. The stated purpose of the format change is removing stitching, so a High NA design goes from two half-field masks to one full-field mask. Masks per advanced design fall, which weakens the unit-volume assumption behind merchant mask growth.

The third is carriers. Because the pod geometry changes, every EUV reticle pod, stocker and in-fab transport at every High NA fab gets replaced rather than expanded, at a new price point.

Second-order companies that could be affected

Gudeng Precision (3680.TW) makes reticle boxes, EUV pods and 300mm wafer carriers. Reticle boxes and EUV pods account for roughly 60% of its revenue [6], which on TWD 8.11bn of trailing-twelve-month revenue is about TWD 4.9bn. [7] A 6x12 reticle fits no existing EUV pod, stocker or in-fab reticle transport, so its largest product line could shift from units-follow-wafer-starts to a whole-fleet replacement at a higher ASP, with revenue and gross profit the likely landing points. What still has to be settled is whether the new carrier is a wholly new specification or an adapted 300mm-class container.

Tekscend Photomask (429A.T) is the only listed pure-play merchant photomask maker. It listed in October 2025 and is putting $400-500m of IPO proceeds into EUV-capable mask capacity in Dresden and Tokyo on the 6x6 format; high-end masks at nodes of 28nm and below are over 55% of revenue. [8] FY3/2026 revenue was JPY 129.6bn against a JPY 373bn market capitalization. [4] It could come under pressure: that 6x6 spend now has a 2033 date on the end of its leading-edge window, and removing stitching simultaneously cuts the number of masks per High NA design. The landing points would be the depreciation life and utilization of those assets. The shares fell 5.53% on September 8; Hoya fell 3.40%, AGC 3.84% and Toppan 4.31% the same day. [4]

Lasertec (6920.T) makes actinic mask inspection systems, including the ACTIS A300 series it has just launched for high-NA masks. [9] Those systems' optical stages and plate handling are built around 6x6, so a format change could force a new platform generation across the Semiconductor Related Device segment, which was 80.7% of FY6/2025 revenue. [4] The direction is two-sided: a mandated replacement is an extra upgrade round across the installed base, but it is also the first full requalification since EUV began, which could give KLA, Applied Materials or Zeiss an entry window. Lasertec does not split mask inspection from wafer inspection, so 80.7% is an upper bound.

How to verify this

The most direct checkpoint is SPIE Photomask/BACUS papers and supplier roadmaps through 2027: whether Hoya and AGC commit to a 6x12 blank line, and whether pellicle, writer and inspection vendors publish 6x12 programmes.

At the company level, watch whether Gudeng announces a large-format reticle pod programme or a named joint development; whether Lasertec's briefings from FY6/2027 show an R&D step-up or a new platform; and whether Tekscend's capex disclosure re-scopes the Dresden and Tokyo expansion or shortens depreciation lives on EUV mask assets.

The chain breaks if: ASML or TSMC states 6x6 remains the production format past 2033, or the initiative slips without a date; 6x12 is confined to the captive mask shops at Intel, TSMC and Samsung, leaving merchant makers out of the market; the 6x12 reticle is specified to travel in a 300mm-class carrier existing FOUP suppliers already build; or Lasertec states its optical column is substrate-size agnostic.

On timing, nothing here changes any company's 2026-2028 income statement. Specification and development contracts land between 2027 and 2030 as the pilot line is designed, with the pilot line in 2031 and production readiness in 2033; TSMC's High NA ramp from 2030 runs on 6x6 with stitching in the interim. The observable checkpoints are the annual September SPIE photomask cycle and the first 12-inch tool development orders, not an earnings date.

This is only meant to surface transmission chains you may have overlooked - it is not a stock recommendation.

Sources

[1] ASML press release · 2026-09-08 · https://www.asml.com/en/news/press-releases/2026/tsmc-and-asml-announce-industry-transition-to-large-format-photomasks-for-high-na-euv

[2] GlobeNewswire · 2026-09-08 · https://www.globenewswire.com/news-release/2026/09/08/3357321/0/en/asml-and-tsmc-announce-initiative-to-pioneer-industry-transition-to-large-format-photomasks-for-high-na-euv.html

[3] Yahoo Finance · 2026-09-08 · https://finance.yahoo.com/technology/ai/articles/asml-wins-over-tsmc-samsung-060500993.html

[4] Drillr database · 2026-09-08 · daily closes, market capitalization and reported segment mix

[5] Mark Lapedus (Semiecosystem) · 2026-09-08 · https://marklapedus.substack.com/p/tsmc-samsung-back-high-na-euv-and

[6] Taipei Times · 2025-01-21 · https://www.taipeitimes.com/News/biz/archives/2025/01/21/2003830547

[7] stockanalysis.com · 2026-09-08 · https://stockanalysis.com/quote/tpex/3680/

[8] iamfabian · 2026-06 · https://iamfabian.substack.com/p/the-landscape-of-semiconductor-photomasks

[9] Lasertec product page · 2026-09-08 · https://www.lasertec.co.jp/en/products/semiconductor/actis_a300.html