ASML raises 2026 sales forecast on AI chip demand

ASML raised its 2026 revenue forecast to €36B-€40B from €34B-€39B after Q1 net sales of €8.8B beat the €8.5B consensus and net income of €2.8B topped estimates. Management said AI-driven chip demand is outpacing supply, and guided Q2 net sales of €8.4B-€9.0B with gross margin of 51%-52% and full-year margin of 51%-53%. The outlook is constructive, though China/export-control risk could pressure results toward the low end of guidance.

Analysis

The important read-through is not just that ASML’s order environment is improving, but that capacity spending is becoming more front-loaded into the next 12-18 months. That tends to pull forward revenue for the entire semi capex chain, but it also concentrates upside in the most supply-constrained tools and services, which should keep pricing and utilization elevated even if end-demand normalizes later. The market is likely underestimating how much of this is a “capacity lock-in” trade: once foundry and memory customers commit, they have to keep spending to avoid losing node positioning, which creates a multi-quarter compounding effect.

The second-order winner is TSMC, because its capex budget now looks less discretionary and more like a defensive response to AI scarcity; that should support equipment demand while also reinforcing TSMC’s pricing power versus smaller foundries. Memory is the more interesting marginal signal: a sharp re-acceleration there usually precedes broader semi cycle tightening, which can later lift DRAM/NAND pricing and improve sentiment for SK Hynix/Micron-type exposure even though ASML is the direct beneficiary. By contrast, any incremental China restriction risk is more of a valuation ceiling than a business-model breakage, because displaced demand is likely to reappear elsewhere while the global tool shortage persists.

The contrarian point is that consensus may be too focused on export-control downside and not enough on the fact that ASML’s book is now being supported by both AI logic and memory, reducing dependence on one end market. The bigger near-term risk is not demand, but whether supply chain bottlenecks or regulatory frictions delay shipments and push revenue recognition beyond the market’s patience window. That creates a classic setup where the stock can still rerate higher over months, but becomes vulnerable to any quarter where shipment timing, not demand, becomes the headline.

From a portfolio perspective, this argues for staying constructive on the EUV supply chain while fading the idea that ASML is a clean “beat-and-raise forever” story; the upside is durable, but the path will likely be noisy around policy headlines and delivery cadence. The strongest setup is a relative-value expression versus lower-quality semi equipment names that do not have ASML’s scarcity moat or backlog visibility.

Key Decisions for Investors

• Add to ASML on 1-2 week post-earnings weakness, targeting a 3-6 month hold; the setup is favorable as long as gross margin guidance stays anchored above 51%, with upside driven by backlog conversion rather than multiple expansion alone.
• Overweight TSM vs. SMH/sector ETF for the next 3-6 months; TSMC should be the cleaner beneficiary of forced capex intensity and has better visibility into AI-led wafer starts. Use a 1x1 pair if you want to isolate the demand signal from broader semiconductor beta.
• Initiate a tactical long in memory exposure via MU or SK Hynix proxy baskets on any 5%+ pullback over the next 2-8 weeks; the memory mix shift suggests the cycle is tightening earlier than consensus models imply.
• Avoid shorting ASML on China/export-control headlines unless policy becomes binding; the better expression is to short weaker equipment names with less pricing power and lower backlog quality, since demand displacement can cushion ASML more than the market expects.
• If you want convexity, buy 6-9 month ASML call spreads rather than outright equity; this captures the medium-term rerating from capex lock-in while limiting downside if shipment timing gets pushed by regulatory noise.