ASML faces a double-edged Terafab: prospective customer and potential EUV rival
Elon Musk's Terafab project presents ASML with an unusual dilemma: the Texas mega-fab could become one of the largest single buyers of ASML's EUV lithography machines — or it could try to bypass ASML entirely.
The customer case. Terafab, announced in March 2026 as a joint venture between SpaceX , Tesla , xAI and Intel, targets 2nm chips at a scale of 1 terawatt of annual AI compute output. A fab at that node cannot run without ASML's EUV tools, and ASML is the sole global supplier. On 20 May 2026, ASML CEO Christophe Fouquet confirmed direct talks with Musk, calling him "very serious" but cautioning that ASML's EUV backlog — already stretched by TSMC, Samsung and Intel — could be the binding constraint. ASML reported Q1 2026 net sales of €8.8 billion and full-year guidance of €36–40 billion. No equipment order has been announced; the relationship remains a CEO-level supply discussion.
The competitor case. In August 2026, Musk signalled interest in free-electron laser (FEL) lithography — a fundamentally different EUV light source that could replace the laser-produced plasma (LPP) sources inside ASML's machines. When X user @beffjezos posted that FEL lithography was "confirmed" and "insanely bullish for Terafab," Musk replied simply: "FEL FTW." A FEL-based source uses a particle accelerator to generate high-power EUV light, offering advantages over ASML's LPP approach: higher power output, no tin debris, wavelength tunability, and potentially lower operating costs. Critically, it could allow Terafab to build its own EUV infrastructure rather than depend on ASML's supply-constrained tools.
FEL-EUV is not yet production-ready. Startup xLight has received $150 million in US CHIPS Act incentives from NIST to build a prototype FEL light source for lithography, and Japanese researchers at KEK are exploring a similar energy-recovery-linac approach. Both remain at the experimental stage. But if the technology matures, it would be the first credible alternative to ASML's EUV monopoly — and Musk's Terafab, with its vertically integrated ambitions and deep capital reserves, is positioned to be an early adopter.
For ASML, the strategic question is stark: sell EUV machines to Terafab and help enable a massive new chipmaking rival to its existing customers, or risk watching Terafab develop an alternative that could erode the monopoly ASML has spent decades building.
