ASML Sees Up to 110 EUV Machines in 2028 Amid Strong AI Demand
ASML, a leading supplier of extreme-ultraviolet (EUV) lithography machines for chip manufacturing, is planning to produce up to 110 such machines by 2028. This comes as JPMorgan highlights 'very very strong' demand for artificial intelligence (AI). ASML's shares fell after CEO Dario Amodei and others called for a slowdown in AI capability growth. The main constraint now is assembly speed, not…
Key points
- ASML plans up to 110 EUV machines for 2028
- JPMorgan stresses strong AI demand
- Main constraint is assembly speed
ASML Eyes 110+ EUV Machines as JPM Stresses 'Strong' AI Demand
bing.com · 14 September 2026
Benzinga and Yahoo Finance LLC may earn commission or revenue on some items through the links below.
ASML Holding N.V. (NASDAQ:ASML) is exploring ways to produce more than 110 extreme-ultraviolet (EUV) chip-printing machines in 2028 as artificial-intelligence demand remains "very very strong," JPMorgan says.
ASML's Amsterdam-listed shares fell 5.2% Monday as chip stocks sold off after Anthropic CEO Dario Amodei urged the industry to slow frontier AI capability gains. Sam Altman and Elon Musk also backed the proposal.
Customers Are Already Booking 2028 Machines
JPMorgan analysts met ASML CFO Roger Dassen, according to Reuters, and said most new EUV orders are now for machines scheduled for delivery in 2028.
ASML said in July that its 2026 low-NA EUV capacity was around 65 machines. It plans to increase that by 30% in 2027 and has been examining another 30% increase for 2028.
ASML is already nearly sold out for 2027, when it can produce at least 80 machines. JPMorgan said ASML's main constraint is now assembly speed rather than its supply chain.
The bank expects Intel Corp. (NASDAQ:INTC) to become a "meaningful customer again" in 2027. Musk's planned Terafab is also "developing" as a customer.
Why ASML Matters to Nvidia
ASML makes machines that chip manufacturers use to print microscopic circuit patterns onto silicon wafers. Chipmakers use EUV technology to produce many of the world's most advanced processors.
Its customers include Taiwan Semiconductor Manufacturing Co. (NYSE:TSM), Samsung Electronics and Intel. Nvidia Corp. (NASDAQ:NVDA) relies on foundries including TSMC to manufacture its GPUs using equipment supplied by companies such as ASML.
ASML is effectively the only commercial supplier of EUV lithography machines, which cost roughly $200 million each.
Prediction Traders Still Doubt an AI Bust
Polymarket gives an AI industry downturn by Dec. 31 a 13% chance, with about $2.4 million traded on the contract. It's trading slightly higher this morning, up from 11% last week.
The market resolves Yes only if at least three listed conditions occur within 90 days. One condition is a 50% fall from an all-time high for ASML or another major AI hardware supplier.
The selloff suggests investors are questioning how quickly AI spending can keep growing. ASML's order book shows companies building advanced chip capacity are still planning years ahead.
Image: Shutterstock
This article ASML Eyes 110+ EUV Machines as JPM Stresses 'Strong' AI Demand originally appeared on Benzinga.com
© 2026 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.
This text was published by bing.com and written by Daragh Thomas. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us.
More in Enterprise & Industry
All →- Human-in-the-loop oversight crucial for enterprise AI: Experts Explain · 1 src
- Crusoe Opens Second Tulsa Plant for AI Components · 1 src
- Meta Stock Could Reach $900 by 2027: Analyst · 1 src
- Buildots secures $130M to expand AI-driven construction monitoring for data centers · 1 src
- Magnitude's Rami Habal on Autonomous Third-Party Risk Management · 1 src
Comments
via GitHub Discussions