ASML Solidifies Chipmaking Monopoly as TSMC and Samsung Adopt High-NA EUV
The world's leading foundries are committing to ASML's next-generation lithography to meet the surging demand for advanced AI hardware.
ASML is strengthening its absolute grip on the semiconductor industry as the world's leading foundries commit to its next-generation High-NA EUV lithography machines. The adoption of this technology by Intel, TSMC, and Samsung ensures that the roadmap for the most advanced AI chips remains dependent on a single supplier.
While Intel acted as the early adopter, TSMC and Samsung have now committed to the technology to support sub-2nm nodes and the escalating requirements of AI chip production. These High-NA (High Numerical Aperture) machines carry a staggering price tag of approximately $400 million per unit. The systems allow manufacturers to print smaller, more complex circuit patterns on silicon wafers, providing the precision necessary for the next leap in transistor density.
The Evolution of Lithography
ASML currently holds a total monopoly on Extreme Ultraviolet (EUV) lithography, the essential process for producing any chip below 7nm. High-NA is the next critical evolution of this hardware, increasing the numerical aperture from 0.33 to 0.55. This technical shift enables significantly higher resolution, allowing chipmakers to avoid "multi-patterning"—the inefficient process of using multiple exposures to create a single layer. By eliminating these extra steps, manufacturers can reduce both production time and overall complexity.
Implications for AI Hardware
This transition is critical for the future of artificial intelligence. As AI models grow in complexity, the industry requires transistors with higher density and superior power efficiency to maintain performance gains. Because ASML is the sole provider of both standard EUV and High-NA EUV machines, the global semiconductor ecosystem is locked into a single point of failure. Any delivery delays or technical hurdles at ASML directly impact the production timelines for the world's most advanced logic and memory chips.
The Path Forward
Industry observers are now watching how quickly these systems can be scaled across the three major foundries. While the commitment from TSMC and Samsung signals a broad industry consensus on the necessity of High-NA, the high cost of entry further concentrates power among a few elite players. The primary focus remains on whether the technical success of these machines can keep pace with the aggressive scaling targets set for the next generation of AI hardware.