High-NA EUV adoption timelines are splitting across logic and memory.
$INTC went first — limited production on select Panther Lake layers as of July 2026. $TSM is holding current EUV through 2029 and won't ramp High-NA volume until 2030. $SSNLF is prioritizing DRAM over foundry, targeting 2028. $SKHY follows the same 2028 DRAM window.
ASML's bigger-mirror scanner prints finer features in fewer steps, but rollout is staggered by use case. The shift from 6-inch to 12-inch photomasks is also in motion to boost throughput.
Intel's early move buys them a headline, but volume ramps tell the real story. TSMC's delay signals they're squeezing more out of existing tools before committing capex. Samsung's memory-first approach reflects where margin pressure sits.
Roadmap dates ≠ sold-out capacity. Watch who converts slides into actual wafer starts.
$INTC went first — limited production on select Panther Lake layers as of July 2026. $TSM is holding current EUV through 2029 and won't ramp High-NA volume until 2030. $SSNLF is prioritizing DRAM over foundry, targeting 2028. $SKHY follows the same 2028 DRAM window.
ASML's bigger-mirror scanner prints finer features in fewer steps, but rollout is staggered by use case. The shift from 6-inch to 12-inch photomasks is also in motion to boost throughput.
Intel's early move buys them a headline, but volume ramps tell the real story. TSMC's delay signals they're squeezing more out of existing tools before committing capex. Samsung's memory-first approach reflects where margin pressure sits.
Roadmap dates ≠ sold-out capacity. Watch who converts slides into actual wafer starts.
