Not Sandisk. Not Micron. This AI Semiconductor Powerhouse Could Be the Biggest Winner of the Memory Boom

ASML expects memory-related net system sales to rise 75% in 2026, as HBM/DRAM capacity ramps to meet AI data-center demand. Memory manufacturers’ capex is projected to jump 67% to $97B this year and then rise another 50% to $146B in 2027, supporting continued equipment spend and likely margin expansion from ASML’s EUV lithography monopoly. Overall, the article is bullish on ASML outperforming peers like Micron and SanDisk due to stronger pricing power.

Analysis

ASML is the cleaner expression of the memory upcycle because it monetizes the capex decision rather than the chip price itself. If HBM/DRAM tightness persists, the first derivative winner is the tool supplier with the highest switching costs and the least customer concentration risk; that shifts returns from cyclical gross margin beta to something closer to quasi-monopoly pricing power. The market often underestimates how much of a memory boom leaks into equipment demand before wafer starts translate into chip revenue.

Second-order, the beneficiary set is broader than MU: every incremental dollar of memory capex pulls through to lithography, metrology, and deposition vendors, but ASML should capture the highest quality of that spend because EUV is the bottleneck node for advanced memory architectures. By contrast, MU and the memory makers may see operating leverage, but they are also the ones forced to reinvest aggressively, so free cash flow can lag earnings and buyback capacity can get crowded out. That makes ASML a better way to own the cycle if the thesis is duration rather than spot pricing.

The key risk is that the market is extrapolating a multi-year shortage linearly. If AI data-center buildouts normalize, HBM supply catches up faster than expected, or customers pause orders after securing multi-year contracts, tool demand can gap down quickly even while end-demand remains healthy. In that scenario the trade works for months, not years: ASML can rerate on visibility, but MU can de-rate on capex intensity and inventory digestion. The contrarian view is that the move may be underdone in ASML relative to the durability of the bottleneck, but overdone in assuming every memory dollar translates into equally durable tool demand.

Key Decisions for Investors

• Long ASML on any post-print pullback; target a 6-12 month horizon where order visibility and margin mix can re-rate the stock. Falsifier: memory capex guidance rolls over or ASML order intake decelerates for two consecutive quarters.
• Pair trade: long ASML / short MU into strength over the next 1-3 months. Thesis is that ASML captures the toll-booth economics while MU bears the reinvestment burden. Exit if MU raises long-term capex guidance without margin dilution.
• If implied vol is reasonable, buy ASML 6-9 month call spreads instead of stock to express upside from memory capex acceleration with defined risk. The trade works best if the market has not already priced a 2027 capex step-up.
• Watch for evidence that HBM supply is normalizing faster than expected; if lead times compress or customer prebuying eases, reduce ASML exposure because the second-order equipment cycle can inflect before earnings do.