ASML lifted its sales forecast for 2026 and signaled plans to build more lithography systems. The move tracks directly with continued heavy spending on AI training clusters. Nvidia CEO Jensen Huang separately committed an extra $2 billion to European projects on the same day.
Context
The updates surfaced during ASML’s quarterly earnings discussion. The Dutch company makes the extreme-ultraviolet machines that pattern the most advanced chips. Demand for those machines has risen with the scale of AI data-center builds. Nvidia’s additional European pledge arrived at the same time and points to further capital flowing into the same supply chain.
ASML had already been running its factories at high utilization. The new forecast shows management now expects that pace to continue or accelerate through the middle of the decade. The Bloomberg report framed both announcements as responses to the same driver: sustained orders for hardware that supports large language models and inference workloads.
Detail
Bloomberg’s Ed Ludlow walked through the ASML numbers on air. The company tied the higher revenue target to customers ordering equipment for advanced semiconductors used in AI. Production increases will center on systems that can support the throughput required for both training runs and large-scale inference clusters.
Huang’s $2 billion commitment covers fresh European investments on top of Nvidia’s existing activity in the region. The earnings call and the Nvidia announcement together indicate that capital spending on AI infrastructure is expected to remain elevated at least through 2026. No new machine-specific shipment targets or regional revenue splits were released in the segment.
The two pieces of news share a common constraint. Every advanced AI accelerator ultimately requires wafers processed on ASML’s latest EUV and high-NA tools. Increased European spending from Nvidia therefore still depends on ASML delivering more of those tools on schedule.
Why it matters
Engineers and technical founders planning multi-year hardware roadmaps now face a clearer picture of where the real limit lies. The raised forecast does not remove the bottleneck; it only confirms that demand has outrun prior supply assumptions. If ASML misses its production ramp, lead times for the highest-end lithography equipment will stay long and prices will remain elevated.
Nvidia’s European outlay may speed up local data-center construction, yet the money ultimately competes for the same scarce machines. Teams that need guaranteed access to new silicon in 2026 must now track ASML’s factory output and supplier yields rather than assume demand alone will solve the shortage. Any shortfall will keep the cost of frontier training runs higher than software roadmaps currently project.
The announcements also highlight a geographic shift in spending without a corresponding shift in manufacturing capacity. Europe gains incremental investment, but the critical production equipment still comes from a single supplier whose output schedule has not changed. That mismatch will continue to shape where and when new AI capacity actually comes online.
---
Sources:
No comments yet