Intel’s EUV Regret and the High-NA Bet It Can’t Afford to Miss

For most of the semiconductor era, Intel believed manufacturing leadership was a matter of engineering will. If the physics became difficult, Intel’s engineers would find a way around it. That confidence sustained decades of dominance—and, in the late 2010s, produced one of the most costly misjudgments in the industry’s modern history. When extreme ultraviolet lithography, known as EUV, began moving from experimental tool to production necessity, Intel hesitated. Rivals moved forward. The gap that followed reshaped the balance of power in global chipmaking. Today, Intel is determined not to repeat that mistake. Its aggressive push into High-NA EUV—the next generation of lithography—marks a deliberate reversal of philosophy: better to risk being early than suffer the cost of waiting again. A technology Intel chose to wait out EUV lithography uses light with a wavelength of 13.5 nanometers, enabling far finer patterning than deep ultraviolet tools. For years, EUV systems were unreliable, slow, and enormously expensive. Intel’s leadership judged the technology immature and instead pushed deep-UV lithography to its limits through complex multi-patterning. That strategy delayed EUV adoption by several years. By 2019, competitors were already producing chips using EUV at the 7-nanometer class. Intel did not deploy EUV in high-volume manufacturing until Intel 4, which entered production in 2023—roughly four years behind the industry’s leading edge. The cost of that delay was not abstract. Intel’s long-delayed 10nm process, later rebranded as Intel 7, suffered from yield issues and repeated schedule slips. During the same period, rivals steadily refined EUV-based production, compounding their manufacturing advantage. The market did not wait As Intel struggled, TSMC made EUV a core part of its roadmap. By the early 2020s, TSMC was producing advanced chips at scale for Apple, AMD, and Nvidia. Those customers were less interested in who invented the technology than in who could deliver it reliably, on time, and in volume. By 2022, TSMC controlled more than 90% of leading-edge foundry capacity, according to industry estimates. Intel, once the benchmark for manufacturing excellence, had ceded that position. The lesson was painful but clear: engineering ingenuity could not substitute for timely adoption when physics tightened its constraints. Why TSMC is not rushing High-NA TSMC Arizona Factory Intel’s renewed urgency stands in contrast to TSMC’s deliberate caution with High-NA EUV, the next major lithography leap. High-NA tools, which increase numerical aperture from 0.33 to 0.55, promise finer resolution but come with steep trade-offs. Each High-NA EUV machine costs more than $350 million, nearly double the price of current EUV systems. Throughput is lower. Tool availability is limited. Integrating High-NA requires new fab layouts, new resist chemistries, and new process flows. For TSMC, which manufactures chips for nearly every major fabless designer, early adoption carries amplified risk. A misstep does not delay a single company’s product—it can disrupt iPhone launches, AI accelerator roadmaps, and global supply chains. TSMC’s business is built on predictability. With its 2-nanometer process already fully booked years in advance, the company has little incentive to absorb first-mover risk before High-NA economics improve. Waiting is not a technological weakness; it is a reflection of scale and customer obligation. Intel’s different calculus ASML’s High NA EUV Tool at Intel Oregon Factory Intel’s position is fundamentally different. After losing process leadership once, the company sees hesitation as the greater danger. Intel was the first company to receive a High-NA EUV system from ASML, the sole supplier capable of producing such tools. Intel plans to introduce High-NA selectively in its 14-angstrom-class (14A) process later this decade, following its 18A node. The financial commitment is substantial. A single advanced fab can exceed $20 billion in capital expenditure. High-NA tools alone represent billions in upfront investment. But Intel’s leadership has concluded that absorbing early pain is preferable to surrendering another learning cycle to competitors. A lesson written in silicon Intel’s EUV delay will likely be remembered as a case study in how quickly manufacturing leadership can erode when technology inflects. The company’s High-NA push is an attempt to apply that lesson decisively. Whether the gamble succeeds remains uncertain. High-NA EUV is expensive, complex, and unproven at scale. But the strategic shift is unmistakable. In an industry where advantages compound over time, Intel no longer wants to be the company that waits for certainty—only to discover that certainty arrived too late.

Khi "Nhà" Buộc Phải Thực Dụng Để Sinh Tồn

Superhero Fatigue Is Real. And It Might Be Warhammer 40K’s Moment.

The $15 Watch at the End of History

Mọi thứ là một giấc mơ....

If Red Bull Succeeds With Its Engine, Cars May No Longer Be Out of the Question

Có những người không rời đi ngay lập tức