Will Musk Challenge ASML's EUV Monopoly? TeraFab Reportedly Exploring FEL Light Sources for Chip Manufacturing

Tesla (TSLA-US) CEO Elon Musk's reportedly筹备ing wafer fab project 'TeraFab' may abandon ASML's (ASML-US) exclusively monopolized extreme ultraviolet (EUV) laser-produced plasma (LPP) light source technology, opting instead for the free electron laser (FEL) approach to generate the short-wavelength light required for chip patterning.

If true, this would mark the first serious challenge to the core light source technology underpinning advanced semiconductor processes in decades, shaking not only ASML's market position but potentially reshaping the foundry industry's long-standing reliance on a single supplier.

This speculation arose after analysts observed publicly available facility designs for TeraFab and noted that Musk himself did not deny the direction in a subsequent response, causing the discussion to rapidly gain traction within the industry.

(Image: Musk on X)

However, as of now, Musk and his team have not officially confirmed these plans, and all information remains external interpretation.

Analysts point out that this speculation is being taken seriously because FEL genuinely demonstrates performance advantages over LPP that are difficult to match.

Current ASML EUV light sources operate at a fixed wavelength of 13.5 nanometers, with output power long capped around 1 kilowatt. In contrast, FEL not only allows wavelength tuning via electron beam energy and magnetic field settings, but can also easily achieve output powers in the multi-kilowatt to tens of kilowatts range, with coherence and spectral purity far surpassing LPP.

The energy efficiency gap is equally stark: LPP systems require approximately 4.4 megawatts of electricity to produce 1 kilowatt of usable EUV light, resulting in an overall efficiency of only about 0.05%. In contrast, FEL systems using energy recovery technology (ERL-FEL) are estimated to require only 0.7 megawatts to produce the same output, achieving roughly six times higher efficiency.

In fact, pursuing the FEL route is not Musk's original idea. U.S. startup xLight, chaired by former Intel CEO Pat Gelsinger, has already invested heavily in this field.

xLight's architecture involves accelerating electrons in a particle accelerator, then directing them into an undulator with periodic magnetic fields to generate coherent light, which is then distributed from a central accelerator facility to multiple lithography machines within a fab via a fiber-like transmission system.

The company claims its light source can achieve four times the power of current LPP systems, with a single system capable of supplying up to twenty ASML scanners, and a lifespan of up to thirty years, potentially drastically reducing both capital and operational expenditures for wafer fabs.

Notably, despite its performance advantages, FEL still faces a long road before it can truly replace current EUV mass production processes.

Analysts emphasize that EUV lithography demands more than just 'generating light at a specific wavelength'—it requires high stability, high repetition rates, high reliability, extremely low cost, extremely long uptime, and compatibility with existing reflective optical systems. These are precisely the moats ASML has built through years of engineering refinement.

In contrast, FEL requires large-scale equipment such as electron accelerators, undulators, vacuum systems, and beam control systems, resulting in significant footprint and installation costs, making it difficult to replicate ASML's mass production advantages in the short term.

Industry experts suggest FEL's true value may not lie in immediately replacing current EUV, but in paving the way for next-generation lithography at even shorter wavelengths.

As processes advance into the 6nm, 5nm, and sub-5nm regimes, FEL's tunable wavelength, high coherence, and high peak brightness will become increasingly attractive. However, shorter wavelengths also bring heightened challenges for optics, photomasks, photoresists, and vacuum systems.

Besides FEL, other technical routes such as high-harmonic generation (HHG), discharge-produced plasma (DPP), and synchrotron radiation are also under discussion, each with trade-offs in maturity, brightness, and size.

Regardless of which ultimately prevails, all candidate technologies—including FEL—currently face significant engineering hurdles, and their commercialization timelines remain uncertain.

Which light source path TeraFab under Musk will ultimately take remains to be officially confirmed.