Samsung Delays High-NA EUV to 2030, SK Hynix Expands Roadmap in Parallel
Nashnova编辑部
Samsung is pushing High-NA EUV mass production to 2030, aligning it with its 1nm node after the technology failed to land at 2nm and 1.4nm as planned; SK Hynix, meanwhile, has already taken delivery of equipment and will begin R&D in 2026 — a split that exposes deep uncertainty over when this technology can actually make money.
What is High-NA EUV, and why is Samsung delaying it?
High-NA EUV — a next-generation chip lithography system that widens the lens aperture to print finer circuits — raises the numerical aperture from 0.33 to 0.55, improving resolution and potentially cutting the number of repeat-exposure steps.
Samsung had planned to introduce it at the 2nm and 1.4nm nodes. Both missed their windows because supporting materials — photomasks and pellicles — are not ready.
This means → High-NA EUV's first mass-production use is now penciled in for around 2030, tied to the 1nm node. In plain terms = the machine has shipped for years, but the "consumables" can't keep up, so the line can't run.
What does Samsung's production timeline look like now?
The 2nm node is already commercialized — Samsung's most advanced process in production today.
1.4nm (SF1.4) is targeted for mass production in 2029; the enhanced SF1.4+ and 1nm nodes are expected to follow in 2030.
High-NA EUV's timeline is pinned to 1nm. This reflects Samsung's framing of the technology as a gate pass for the sub-1nm era, not a tool to speed up today's nodes.
Where does TSMC stand on this?
TSMC said in 2025 that its A14 process and subsequent enhanced versions do not need High-NA EUV to manufacture.
Its stated position: adopt it only when it delivers quantifiable benefits.
This means → the world's two largest logic foundries — Samsung and TSMC — are both pumping the brakes on High-NA EUV. The bottleneck is industry-wide maturity, not a single company's shortfall.
Why is SK Hynix moving in the opposite direction?
SK Hynix took delivery of High-NA EUV equipment in H2 2025 and plans to begin R&D in 2026 — years ahead of Samsung's mass-production timeline.
At the same time, it has started using phase-shift masks (PSM) — a technique that exploits phase differences in reflected light to sharpen exposure contrast — to squeeze more performance out of existing EUV.
In plain terms = SK Hynix is running a dual track: "max out the old tech's upgrades while getting a head start on the new one." That is the opposite of Samsung's "wait until the ecosystem is ready" approach.
What does this split really tell us?
Samsung's delay + TSMC's caution = the industry consensus is that High-NA EUV is not ready for volume production. The bottleneck sits in materials and process, not in the equipment itself.
SK Hynix's early move is a bet: if supporting materials mature faster than expected, it could convert that head start into a process-technology lead.
This reflects a deeper signal: ASML has been shipping these machines for years, but the distance from "it works in a lab" to "it prints money in a fab" is far longer than the market originally assumed.
Content is for reference only, not financial advice.