xLight vs. ASML: Can Free-Electron Lasers Redefine EUV Lithography?
Extreme Ultraviolet (EUV) lithography has become one of the most strategically important technologies in modern semiconductor manufacturing. Today, ASML dominates the market for advanced lithography systems, while its proprietary Laser-Produced Plasma (LPP) light source remains one of the most complex and expensive components inside every EUV scanner.
A new challenger is attempting to disrupt this model.
Founded in 2021, xLight is developing a particle accelerator-based Free-Electron Laser (FEL) capable of replacing conventional plasma-based EUV sources. Led by former Intel CEO Pat Gelsinger as Executive Chairman, the company aims to introduce a scalable, centralized light-generation architecture that could fundamentally change how semiconductor fabs deploy High-NA EUV tools.
If successful, xLight could become the first major American supplier of next-generation EUV light sources by its planned prototype milestone in 2028.
[Lithography Disruption]
ASML LPP (Laser-Produced Plasma)
│
└── 600W–1kW dedicated light source per scanner
xLight FEL (Free-Electron Laser)
│
└── ~5kW centralized source shared by 20+ scanners
💰 Funding, Government Support, and Growth Strategy #
Unlike many semiconductor startups, xLight has rapidly secured both private investment and government backing.
Key funding milestones include:
- June 2026: The U.S. Department of Commerce committed $150 million through an equity investment under the CHIPS and Science Act.
- Current fundraising: xLight is reportedly pursuing another $350 million round with participation from major semiconductor companies, including ASML, TSMC, Intel, and Micron.
- Capital runway: Combined with its previous Series B financing and non-binding infrastructure funding, the company has secured approximately $550 million in core equity financing while positioning itself for future multi-billion-dollar manufacturing projects.
Rather than subsidizing operations, the federal investment gives the U.S. government direct ownership exposure, highlighting the strategic importance of domestic lithography technologies.
⚙️ Engineering Showdown: LPP vs. Free-Electron Laser #
A modern EUV scanner contains well over 100,000 individual components, with the light source accounting for roughly 15% of the system’s total cost.
Although both approaches ultimately generate EUV photons for semiconductor patterning, their underlying physics differ dramatically.
ASML’s Laser-Produced Plasma (LPP) #
ASML’s existing technology, originally developed by Cymer, operates by creating plasma from molten tin droplets.
The process consists of three primary stages:
- A high-power CO₂ laser strikes molten tin droplets approximately 50,000 times per second.
- The resulting plasma reaches temperatures approaching 500,000°C.
- The plasma emits 13.5 nm EUV light, which is collected and directed into the lithography optics.
This design has enabled commercial EUV manufacturing, but several physical limitations remain:
- Extremely poor wall-plug efficiency.
- Tin debris contaminates collector mirrors.
- Increasing optical power becomes progressively more difficult.
Although ASML has demonstrated experimental 1 kW LPP output, scaling significantly beyond this range presents substantial engineering challenges.
xLight’s Free-Electron Laser (FEL) #
Instead of generating plasma, xLight accelerates electrons to nearly the speed of light.
Its proposed architecture consists of:
- A compact 25-meter linear accelerator (linac).
- Electron beams passing through precision undulator magnets.
- Highly coherent photon emission without vaporizing any material.
Unlike plasma-based systems, FEL output can be electronically tuned by adjusting beam energy and magnetic field strength.
Potential operating wavelengths extend into the 2–7 nm “Blue-X” region, approaching soft X-ray frequencies and offering significantly greater flexibility for future lithography nodes.
+--------------------------------------------------------------+
| xLight Centralized FEL Architecture |
| |
| [25m Linear Accelerator] --> [Undulator Magnets] |
| │ |
| ┌──────────────┬─────────┴──────────────┐ |
| ▼ ▼ ▼ |
| Scanner 1 Scanner 2 Scanner N |
| (20+ scanners supported) |
+--------------------------------------------------------------+
Centralized Light Generation #
Perhaps the most significant architectural difference lies in deployment.
Today’s EUV systems require each lithography scanner to include its own dedicated light source.
xLight instead proposes a centralized facility capable of supplying approximately 5 kW of EUV power to more than 20 scanners simultaneously.
According to the company, this design could:
- Increase throughput by approximately 50% in existing fabs.
- Potentially double productivity in purpose-built facilities.
- Simplify future scalability as EUV power requirements continue increasing.
👥 Leadership and Technical Expertise #
Commercializing accelerator physics requires expertise spanning semiconductor manufacturing, national laboratories, and large-scale engineering.
xLight’s leadership reflects this interdisciplinary approach.
Nicholas Kelez #
Serving as both CEO and CTO, Nicholas Kelez previously led engineering efforts for the Linac Coherent Light Source (LCLS) at SLAC and later worked at quantum computing startup PsiQuantum.
Pat Gelsinger #
Following his departure from Intel, Pat Gelsinger joined xLight in March 2025 as Executive Chairman.
His experience spans semiconductor manufacturing, corporate leadership, supply chain strategy, and U.S. semiconductor policy initiatives.
Technical Advisors #
The advisory team includes industry veterans such as:
- Jim Wiley, former ASML executive specializing in EUV infrastructure.
- Sanjay Natarajan, former Intel Senior Vice President with more than three decades of semiconductor manufacturing experience.
Research collaborations also extend to:
- Cornell University
- Los Alamos National Laboratory (LANL)
- Fermilab
🏭 A “Parasitic Innovation” Strategy #
Rather than competing directly with ASML by building complete lithography systems, xLight focuses exclusively on replacing one subsystem—the EUV light source.
xLight Strategy
│
┌─────────────┴─────────────┐
▼ ▼
Strategic Advantages Structural Risks
• Focused subsystem R&D • Dependent on ASML adoption
• Avoids optics development • Export control uncertainty
• Avoids scanner integration • Competition from ASML R&D
This approach significantly reduces development complexity by avoiding:
- Projection optics
- Precision wafer stages
- Metrology systems
- Scanner integration
Instead, xLight aims to become a specialized supplier whose FEL modules could eventually replace conventional LPP sources.
However, this strategy also introduces dependencies.
If ASML continues improving its in-house plasma technology, the commercial opportunity for FEL adoption could narrow substantially.
🌍 Strategic and Geopolitical Significance #
Beyond engineering, xLight reflects broader geopolitical priorities.
ASML’s dominant market position effectively makes it a single point of failure within the global semiconductor ecosystem.
Any disruption involving:
- Export controls
- Regional political changes
- Manufacturing restrictions
- Supply chain interruptions
could affect worldwide chip production.
Supporting domestic alternatives therefore aligns with broader U.S. industrial policy aimed at strengthening long-term semiconductor independence.
🚀 Alternative Technologies Targeting 2028 #
xLight is not the only company exploring next-generation lithography.
Several competing approaches are under active development.
X-Ray Lithography #
Backed by Founders Fund, another startup is pursuing complete X-ray lithography systems, including custom scanners and dedicated foundries targeting sub-2 nm manufacturing.
Laser-Induced Discharge Plasma (LDP) #
Several research institutions continue exploring electrical discharge methods to generate EUV radiation with potentially simpler hardware than conventional LPP systems.
Other Emerging Technologies #
Additional candidates include:
- Nanoimprint Lithography (NIL): Promising for memory manufacturing but currently limited by overlay accuracy.
- Directed Self-Assembly (DSA): Suitable for repetitive structures but less effective for complex logic layouts.
- Multi-Beam Electron Beam Lithography (MEBL): Offers exceptional precision but lacks the throughput required for high-volume semiconductor production.
🔬 Remaining Engineering Challenges #
Although Free-Electron Lasers have been proven in scientific research facilities, adapting the technology for commercial semiconductor fabs presents significant obstacles.
Major engineering challenges include:
- Miniaturizing accelerator systems into compact 25-meter cleanroom installations.
- Achieving continuous industrial reliability.
- Integrating superconducting RF cavities and energy recovery systems.
- Maintaining beam stability suitable for high-volume manufacturing.
Even if prototype demonstrations at the Albany Nanotech Complex succeed in 2028, widespread adoption will depend on convincing conservative semiconductor manufacturers to redesign factory infrastructure around centralized EUV generation.
📌 Conclusion #
xLight represents one of the most ambitious attempts to reshape the future of semiconductor lithography without competing directly against ASML’s complete scanner ecosystem.
Instead of replacing the lithography machine itself, the company targets its most technically demanding subsystem—the EUV light source.
Whether Free-Electron Lasers ultimately replace Laser-Produced Plasma remains uncertain, but the concept introduces compelling advantages in scalability, efficiency, and future wavelength flexibility. If successful, centralized FEL facilities could redefine how advanced semiconductor fabs are designed, particularly as High-NA EUV pushes toward increasingly demanding process nodes.
With prototype demonstrations targeted for 2028, the coming years will determine whether FEL technology evolves from laboratory physics into a foundational pillar of next-generation semiconductor manufacturing.