The Binding Constraint: From EUV Machines to Megawatts

Print 🖨

EIR identifies five quarterly signals—from EUV tool orders to packaging utilization—that best indicate whether AI chip supply can scale through 2030

CALGARY, Alberta (June 2, 2026) — Enverus Intelligence® Research (EIR), a subsidiary of Enverus, the leading energy data analytics platform, has released a report that frames AI compute growth as a manufacturing-throughput question and recommends a quarterly monitoring framework for assessing whether the semiconductor supply chain can support continued scaling through 2030.

The report’s core premise is that sustained expansion depends on incremental capacity additions, especially new extreme ultraviolet (EUV) tool deployments into high-performance computing (HPC) fabrication lines rather than reallocating existing leading-edge capacity already committed across end markets.

EIR highlights the five indicators that matter most each quarter along with key signals to provide a practical read on whether near-term bottlenecks such as advanced packaging and memory supply are easing and whether longer-term constraints tied to EUV tool delivery rates are tightening or loosening.

“Because leading-edge capacity is structurally committed, the most actionable question each quarter is whether incremental tools and downstream components are landing fast enough to expand AI-serving throughput,” said Carson Kearl, report author and senior analyst at EIR. “Our framework focuses on a short list of signals to help gauge whether the supply chain is on track to support the report’s 2030 capacity scenarios.”

Key takeaways:

The AI ecosystem consumed ~20%–25% of global EUV layer passes in 2025 and is expected to rise to ~30% by 2030, with growth driven mainly by added tools rather than reallocation.
EIR’s base case projects AI chip production scaling from ~1.2 GW in 2023 to 25 GW by 2030E, with a scenario range of 18–35 GW.
Near-term constraints remain material in the report’s framing with advanced packaging cited as the main near-term bottleneck.

EIR’s analysis pulls from a variety of products including Enverus ONE.