AI

TSMC and Samsung commit to ASML's $400M chip machines for AI

The foundry giants will deploy High NA EUV lithography by 2030 and join initiative to cut chipmaking costs with larger photomasks.

Omega Editorial· September 8, 2026· 3 min read

TSMC and Samsung have committed to deploying ASML's most advanced chipmaking technology, joining an industry push to manufacture the complex semiconductors required for AI workloads while reducing production costs.

Both companies announced Tuesday they will use ASML's High NA extreme ultraviolet lithography machines for advanced chip production. Samsung plans to deploy the technology for DRAM manufacturing by 2028, while TSMC targets high-volume production at advanced nodes starting in 2030, according to statements from the companies first reported by Quartz.

The commitments make TSMC and Samsung the second and third customers for ASML's High NA EUV systems, following Intel. Each machine carries a price tag of approximately $400 million.

Transition to larger photomasks

Alongside the lithography commitments, TSMC and Samsung joined ASML in launching an industry initiative to transition from today's 6-inch photomasks to a 12-inch format. Photomasks are the patterned stencils that transfer circuit designs onto silicon wafers during chip production.

The initiative aims to establish a 12-inch mask pilot line by 2031, with full production readiness expected by 2033. ASML and TSMC said the larger format should increase fabrication productivity, lower chipmaking costs, and eliminate stitching constraints that currently add complexity and expense to production.

"By bringing together expertise from across the industry's value chain, we hope to keep providing the benefits of cutting-edge technology through continuous innovation that lowers barriers and puts advanced solutions accessible at scale," TSMC Chairman and CEO C.C. Wei said in a statement.

Why it matters

The coordinated industry move reflects the escalating technical demands of AI chip production. As AI workloads require more intricate transistor architectures, chipmakers need higher-resolution lithography tools and more efficient manufacturing processes to maintain economic viability at advanced nodes. The $400 million per-machine investment signals how critical—and expensive—staying competitive in AI semiconductor manufacturing has become.

AI demand drives record results

The announcements come as AI-driven semiconductor demand has pushed both companies to record performance levels. TSMC reported July revenue of NT$467.58 billion, up 44.7% year-over-year, and raised its full-year capital expenditure projection to between $60 billion and $64 billion for 2026.

Samsung's semiconductor division has posted record results, with the company's stock crossing a $1 trillion market capitalization earlier this year on surging memory chip demand tied to AI infrastructure spending.

TSMC said that as technology nodes advance, a growing number of chip layers will require High NA EUV, a trend driven by the more complex transistor architectures that AI workloads demand. Samsung, drawing on experience in both memory and foundry manufacturing, said the technology would enable further DRAM scaling by improving resolution and reducing process complexity.

ASML President and CEO Christophe Fouquet said High NA EUV adoption would increase along the device scaling roadmap, first using current 6-inch masks and then supported by 12-inch masks. Samsung also announced an expansion of its existing partnership with ASML covering High NA EUV and advanced semiconductor manufacturing technologies more broadly.

Details were first reported by Quartz.

#semiconductor manufacturing#euv lithography#tsmc#samsung#asml#ai chips

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in AI

AI· 3 min read

U.S. Data Center Capacity Set to Nearly Triple by 2030

Power, land, and cooling infrastructure emerge as the next bottleneck in AI expansion as tech giants race to build computing facilities.

Via AI Watch · Sep 8, 2026
AI· 4 min read

AI in Clinical Trials Has a Judgment Problem, Not a Tech Gap

Industry experts warn that automation bias and pattern recognition are masquerading as clinical reasoning in trial operations.

Via AI Watch · Sep 8, 2026
AI· 3 min read

OpenAI's AI agents now produce 3 days of research per human day

Chief scientist warns monitoring systems are failing as the company races toward recursive self-improvement, raising questions about whether AI development should slow down.

Via AI Watch · Sep 8, 2026