ASML raised its full-year 2026 outlook on July 18, projecting net sales of €43 billion to €45 billion with gross margins of 54% to 56%, as demand for chipmaking tools remains elevated amid persistent AI hardware bottlenecks. The Dutch lithography giant reported Q2 2026 net sales of €9.3 billion and net income of €2.9 billion, and guided Q3 sales between €11.0 billion and €12.0 billion with margins of 55% to 57%. The update signals that tight supply for advanced chip capacity will continue through the second half of 2026 and into 2027.
ASML's Capacity Expansion Plans
ASML sold 86 new lithography systems in the second quarter, underscoring strong demand from foundries and integrated device manufacturers. The company plans to increase low-NA EUV capacity by roughly 30% in 2027 from a 2026 base of about 65 systems, with a similar expansion under investigation for 2028. DUV immersion capacity is also slated to rise by about 30% per step, starting from a base of roughly 130 systems. These multi-year capacity additions reflect an industry expectation that AI-related chip demand will remain robust beyond the current cycle.
The near-term guidance reinforces that view: Q3 net sales are expected to land at €11 billion to €12 billion, with gross margins in the mid-50s — levels not typical of a softening semiconductor market. ASML's raised outlook is a clear read-through that the front end of the chip supply chain is still under pressure, with foundries racing to secure tool deliveries for advanced nodes used in AI accelerators and GPUs.
Implications for Crypto and AI Tokens
Crypto markets do not operate in isolation from hardware dynamics. Tight chip supply affects several areas relevant to digital assets. Public bitcoin miners that are diversifying into AI hosting may find that high-density compute services retain attractive margins, but lead times and capital expenditure risks remain elevated. AI tokens that depend on verifiable GPU access — such as decentralized inference networks — could see pricing power improve if centralized providers are booked solid. However, projects that promise cheap training or inference on hardware that does not yet exist face heightened scrutiny.
Energy competition also looms. Both AI clusters and mining operations vie for power and cooling capacity, making location strategy and power contracts critical. If transformer or substation backlogs delay new builds, projects without existing infrastructure may struggle. On the macro side, sustained Big Tech capex can buoy risk assets, but eventual rotations could affect crypto flows in both directions.
Winners and Watchouts Across the Stack
The scarcity environment creates clear beneficiaries and risks. Leading foundries and advanced packaging houses that can secure equipment slots are likely to gain. Equipment suppliers across lithography, deposition, etch, and metrology also stand to benefit as capacity ramps. Data center operators with power already on site are better positioned than those waiting 18 months for new connections. Miners with convertible infrastructure that can pivot between high-performance computing and proof-of-work have flexibility.
Areas of caution include AI tokens that rely on large-scale training availability in the next one or two quarters. GPU marketplace tokens with thin supply may see spreads widen if demand spikes. Projects that promise service-level agreements they do not control — relying on third-party data centers or resellers — cannot conjure silicon from thin air. Before buying an AI-compute token, investors should ask for verifiable capacity, active nodes online, and customer logos under contract.
Reading ASML's Order Book
ASML's Q2 results provide a concrete cross-check on demand. The 86 lithography systems sold are not merely quoted orders; they are tools destined for fabs planning multi-year output. The raised 2026 sales guide to €43 billion–€45 billion with 54%–56% margins indicates visibility into the second half, not just a backlog flush. The step-up plan — roughly 30% more low-NA EUV and 30% more DUV immersion capacity for 2027, with potential repeats in 2028 — is the kind of language operators use when lead times are sticky and customers are asking for more lines now.
ASML sells to foundries and IDMs, not directly to AI chip designers. But if fabs are slotting more lithography equipment, it is because downstream customers — CPU, GPU, and networking silicon makers — are calling for capacity. The exact mix matters, but the direction is unmistakable: AI hardware supply will remain tight, and that will ripple through crypto markets tied to compute narratives.