"So yes, NXEs are running flat-out to support 5nm AP demand"
Copilot:
You’re spot on to connect your Samsung A15’s 5nm application processor (AP) with EUV lithography capacity — let’s break it down clearly:
What “5nm AP” Means- The application processor in your A15 (likely a low-cost Exynos or MediaTek variant depending on region) is fabbed on a 5nm-class node.
- “5nm” in modern foundry terms (TSMC N5, Samsung 5LPE, etc.) requires EUV lithography for several critical layers.
- That means ASML NXE scanners (EUV tools) are indeed part of the production flow for chips like yours.
NXE EUV Throughput
- NXE systems (e.g., NXE:3400C/D) are the workhorses for 5nm and below.
- Each wafer needs dozens of EUV exposures, so demand for NXE throughput scales directly with smartphone AP demand.
- Foundries like TSMC, Samsung, and Intel have expanded EUV fleets to keep up with volume nodes (5nm, 4nm, 3nm).
Smartphone Demand Link- TrendForce’s report you were reading shows 328M smartphones produced in 3Q25, with Apple and Samsung leading.
- Since both companies rely heavily on EUV-based APs (Apple’s A-series, Samsung’s Exynos, Qualcomm Snapdragon), yes — NXEs are being run hard to meet this seasonal surge.
- The bottleneck TrendForce flagged wasn’t AP supply, but memory (DRAM/NAND). AP production is scaling fine because EUV fleets have matured.
Nuance- “Cranking NXEs” is accurate in the sense that EUV utilization is very high at 5nm and below.
- But EUV tool throughput is now stable enough that AP supply isn’t the limiting factor — memory shortages are the real constraint.
- In other words: your cheap A15 having a 5nm AP is proof EUV has scaled into mass-market phones, not just flagships.
So yes, NXEs are running flat-out to support 5nm AP demand, but the production choke point in Q3/Q4 2025 is memory, not APs.
Mike, since you’re tracking Multibeam, DensityAI, and Cerebras, do you want me to map how EUV/AP scaling at 5nm intersects with wafer-scale disruption? It’s a fascinating contrast: NXEs enabling mass-market chips, while wafer-scale players bypass traditional scaling altogether.
Great catch, Mike — let’s unpack what TrendForce meant in that line, using the surrounding page context:
Supply Issues & Downward Revisions- The report highlights constrained memory supply (specifically DRAM and NAND) as a key risk factor.
- When they say “supply issues may force downward revisions”, they’re warning that component shortages could limit smartphone makers’ ability to meet production targets, even if demand is strong.
- This isn’t about application processors (APs) being in short supply — the page context points squarely at memory (DRAM/NAND) as the bottleneck.
Implications
- APs (application processors): The text doesn’t suggest AP demand is exceeding supply. In fact, AP availability isn’t flagged as a constraint in this report.
- NAND/DRAM: Yes — demand is higher than supply here. Rising prices and tight availability are expected to squeeze margins, especially for entry-level smartphones. That’s why TrendForce cautions that overall 2025 production growth (currently projected at +1.6% YoY) could be revised downward.
Bottom Line- The warning is memory-specific: NAND and DRAM supply tightness could cap production growth.
- AP supply isn’t mentioned as a limiting factor in this context — so the risk is not across the board, but concentrated in memory.
Would you like me to cross-check current DRAM/NAND supply-demand balance (outside this report) to see if the broader semiconductor market is confirming the same tightness trend? That way you’d know if this is a localized smartphone issue or part of a wider memory crunch. |
