Nikon is mounting a direct challenge to ASML’s dominance in semiconductor lithography—not by chasing EUV, but by aggressively pricing its ArF immersion scanners. This move appears technologically conservative but strategically astute, targeting a critical fracture point in today’s chipmaking landscape: constrained EUV availability, yield pressures at 3nm and below, and the urgent push for onshore capacity in the U.S. Under these conditions, cost efficiency in mature and quasi-advanced nodes has reemerged as a decisive factor for foundries.
While ASML commands near-total control of the EUV market (with effectively 100% share), it also dominates the ArF immersion segment that underpins processes from 28nm down to 3nm. Yet the economics are stark: an ASML NXE:3600D EUV tool costs over $200 million, whereas Nikon’s latest NSR-S636E ArFi system is reportedly priced around $50 million—less than a quarter of the EUV alternative. Crucially, Nikon has vertically integrated key components like projection lenses and wafer stages, further reducing costs and promising shorter lead times. For fabs racing to expand capacity amid ASML’s multi-year EUV backlog, this trade-off is increasingly compelling.
Nikon is already in advanced talks with multiple U.S. and Asian chipmakers, with some discussions nearing purchase agreements. Although no customers have been named, Intel stands out as the most plausible early adopter. CEO Pat Gelsinger has repeatedly cited ASML’s delayed EUV deliveries as a bottleneck for Intel 4 and Intel 3 ramp. In 2025, Intel’s new fabs in Oregon and Arizona still rely heavily on DUV tools for metal layers and backend processes. If Nikon can deliver performance close to ASML’s Twinscan NXT:2050i—at significantly lower cost and faster delivery—Intel has strong incentive to dual-source and de-risk its supply chain.
This isn’t merely about swapping machines. Nikon plans to launch a next-generation ArF platform by 2028, featuring a new high-numerical-aperture (NA) lens and an upgraded wafer stage. The timing is deliberate: it targets the final major upgrade window for DUV before High-NA EUV becomes mainstream. Nikon is betting that through 2030, over 40% of advanced logic capacity will still run on ArFi platforms—especially in Taiwan, China, South Korea, and U.S. fabs that lack access to or cannot afford full EUV clusters.
TSMC and Samsung have fully embraced EUV for leading-edge logic, yet their mature nodes (28nm, 16nm) still account for nearly 30% of revenue and are expanding due to automotive, industrial, and IoT demand. If Nikon penetrates these secondary lines, it could reshape the DUV competitive dynamic. Notably, Nikon isn’t attacking EUV; it’s waging price war precisely where ASML feels most secure—its legacy DUV stronghold. That exposes ASML’s vulnerability: while DUV remains highly profitable, R&D focus has shifted overwhelmingly to EUV and High-NA EUV, slowing innovation in conventional platforms and creating an opening.
In my judgment, Nikon’s success hinges not on specs but on securing a flagship customer. If Intel adopts its tools, it would send a powerful signal to SK Hynix, Micron, and even SMIC that ASML is no longer the sole viable option for advanced DUV. That pressure could force ASML to adjust pricing and delivery terms, alleviating a structural bottleneck in global capacity.
The geopolitical implications are equally significant. The U.S. CHIPS Act is funding dozens of new fabs, most focused on mature nodes. Nikon’s low-cost, rapidly deployable ArFi systems could accelerate America’s shift from a design-centric to a manufacturing-capable semiconductor ecosystem. For Chinese mainland firms, even if Nikon tools remain export-controlled, their mere existence weakens ASML’s perceived indispensability, strengthening negotiation leverage.
Ultimately, this quiet contest in lithography reflects a broader industry recalibration: away from relentless scaling and toward economic viability. As Moore’s Law delivers diminishing returns, cost, delivery speed, and supply chain resilience are becoming more critical than resolution alone. Nikon may never dethrone ASML in the EUV era, but it is laying the groundwork for the first viable alternative infrastructure of the post-Moore era. The pivotal question now is this: once a giant like Intel seriously considers a second source, can ASML sustain its near-mythical status as the irreplaceable linchpin of global chipmaking?