ASML Machine Costs: $5M DUV vs $724M EUV (2026 Prices)

ASML’s pricing follows technology generations, with each step representing roughly a 2x increase in cost while delivering proportionally more transistor density and manufacturing capability.

Machine Pricing Architecture

DUV (Deep Ultraviolet): $5-90 million per machine. Legacy technology but still 60% of ASML’s unit volume with 374 machines sold in 2024. These serve mature nodes and markets like automotive and IoT where cutting-edge manufacturing is not required.

EUV (Extreme Ultraviolet): $183-220 million per machine. Only 44 sold in 2024, but accounting for 38% of system sales revenue. Enables 7nm to 3nm production. ASML sold 44 EUV machines last year with a price tag starting at $220 million.

High NA EUV: $350-400 million per machine. ASML’s cutting-edge High-NA extreme ultraviolet chip — as explored in the economics of AI compute infrastructure — making tools will cost around $380 million each, over twice as much as existing Low-NA EUV lithography systems. Target production is 20 units annually by 2028. Enables sub-2nm nodes.

Hyper NA EUV (2030+): The expected cost for Hyper-NA EUV could exceed $724 million, approximately twice the High NA price. This has caused TSMC, Samsung, and Intel to hesitate on adoption timelines.

The Value Proposition

The economics work because each generation does not just cost more but delivers more. The High-NA EUV technology represents a major breakthrough, enabling an improved 8nm imprint resolution compared to 13nm with current Low-NA EUV tools.

This allows chipmakers to produce transistors that are nearly 1.7 times smaller, translating to a threefold increase in transistor density on chips. Customers pay 2x the machine price but potentially recover the cost through higher yields, faster throughput, and reduced mask costs.

This is part of a comprehensive analysis. Read the full analysis on The Business Engineer.

How AI Is Changing This

AI is revolutionizing ASML’s lithography machine pricing through unprecedented computational complexity and precision requirements. The leap from $5M DUV systems to $724M Hyper-NA EUV machines exemplifies this transformation. A concrete example is ASML’s implementation of AI-powered computational lithography in their latest EUV systems, which uses machine learning algorithms to predict and correct for millions of potential optical aberrations in real-time during chip manufacturing. This AI system processes terabytes of imaging data per second, enabling the production of 2-nanometer chips that would be impossible with traditional optical correction methods. The AI algorithms alone require specialized quantum computing partnerships and cost hundreds of millions to develop, directly contributing to the 14,000% price increase. Additionally, the AI systems demand exotic materials like high-purity tin for plasma generation and advanced neural processing units, making each Hyper-NA EUV machine essentially a supercomputer capable of atomic-level precision manufacturing.