Several startups see a big opportunity to build the next generation of ultrafast lasers, a crucial component in high-end chip manufacturing, in Europe. In June 2026, the European Commission proposed legislation and strategies to reinforce Europe’s competitiveness and independence in the semiconductor and AI ecosystems. Now, one startup is breaking ground on a factory in Lithuania in October.
“Sovereignty isn’t an abstract policy word for us—it’s become the reason funding, partners, and attention are available at all,” says Celia Millon, CEO of RayVen Laser GmBH, in Germany. The EU’s proposed “Chips Act 2.0.” would support the broader computer chip supply chain—including design, materials, and processing equipment like ultrafast lasers. Millon says that shift means significant public funding can reach small startups like hers.
Independent market research puts the global ultrafast-laser market at roughly US $2.4 billion in 2025, growing at around 20 percent a year to reach over $10 billion by the early 2030s. It cites semiconductor miniaturization and advanced packaging as leading drivers alongside precision manufacturing and medical applications.
Femtosecond fledglings
Ultrafast lasers—ones that fire extremely short, femto- to picosecond-long pulses of high-energy light—are among the most valuable tools for advanced material processing. They enable highly accurate processing without causing heat damage to the surrounding material. Ultrafast laser processing is essential for advanced chip components that enable faster data transfer. For example, components called interposers—the wiring layers that help connect stacked chiplets in advanced semiconductor packages—require tiny laser-drilled holes or “through-glass vias,” which manufacturers then fill with copper to route electrical signals. “This is like 3D-printing technology but in glass,” says Nikolajus Gavrilinas, CEO and cofounder of Lithuanian laser startup Litilit. Other applications include PCB probe cards and wafer dicing.
In October, Litilit will open a new factory in Vilnius, Lithuania, targeting the manufacture of up to 3,000 femtosecond lasers per year, which would put it at the high end of global production capacity. When demand is higher than usual, Gavrilinas says, “market players look for less mature companies. For young companies like Litilit, it’s a great opportunity.” While academic buyers of femtosecond lasers tolerate expensive and complex designs, industry needs compact and robust lasers that require no maintenance and operate reliably in factory environments.
In 2014, Gavrilinas and his cofounders patented an approach for generating extremely stable one-micron-wavelength laser pulses using just a few pieces of optical fiber. Through this stable initial pulse, modulating noise becomes straightforward. The architecture achieves 20 percent electrical-to-optical efficiency, and the pure fundamental wavelength gets delivered into a high-energy beam. Over the past decade, the company developed the idea into a compact modular technology and invented sustainable production techniques. “Moving from complexity into stability is the hallmark of a mature technology,” Gavrilinas says, and Litilit’s approach has reached that stage.
“We’ve been approached by companies that do things that were never on our radar.” Benjamin Rudin
Millon’s company, RayVen Laser, meanwhile, has pursued another opportunity: ultrafast lasers that operate at the 2-micron wavelength, which hasn’t yet been widely commercialized. At this wavelength, Millon explains, silicon becomes far more transparent than at more traditional wavelengths. Instead of just ablating a surface, the beam can be focused and structured inside the silicon. That opens new manufacturing options for the semiconductor industry—processes that are either more efficient, require fewer steps, or simply weren’t possible before. That includes backside processing, deep-surface structuring, and through-silicon vias that allow layers in 3D chips to talk to each other. The current technology for those processes involves lots of chemicals and water. Doing it with just lasers is simpler, more cost-effective, and creates less waste material.
RayVen is currently building and qualifying its first two product lines at low-volume pilot scale, rather than mass production. Early customers and collaborators include research institutes, university labs, and industrial R&D groups that need a laser source for material-processing experiments. A new injection of EU funding in April boosted the company’s expansion deeper into the semiconductor processing world. They’re now building demonstrator systems with two partner universities.
Millon says she’s witnessed increased demand on several levels: growing interest from customers, a rising focus on this wavelength at scientific conferences, and new manufacturing needs emerging in photonic integrated circuit manufacturing and copackaged optics.
More data, more lasers
Others are more skeptical. Eric Mottay, cofounder of ultrafast laser giant Amplitude Laser in Bordeaux and now an independent consultant, doubts that Europe’s tech sovereignty campaign will have much influence on its own femtosecond laser companies. Microprocessing is an important domain in general, but it comprises many niche applications that are out of the scope for major players, he says. The telecom and datacom industries, however, have a clear market case for femtosecond lasers—and up-and-coming businesses need to demonstrate capacity for very high volumes. Mottay sees “good potential for EU companies for innovation there.”
One such company is Switzerland-based Menhir Photonics. Their focus is on miniature, precision 1.5-micron-wavelength femtosecond lasers. “Our laser serves as a metronome that synchronizes telecom systems,” explains Menhir CTO Benjamin Rudin. Precise synchronization among cellphone towers, for example, becomes increasingly important with the high data bandwidth in 6G systems. It’s also valuable in quantum technologies, precision microwave generation, and high-speed optical switching in AI data centers. “We’ve been approached by companies that do things that were never on our radar,” Rudin says. For example, he’s working with a company that provides data center transceivers for secure transmission.
At the moment, Rudin says, no other company makes ultrafast lasers that have the same combination of specs as Menhir’s. Gain-switched laser diodes can perform a similar function in the telecom world but have worse noise performance and less bandwidth. Compared to traditional academic markets that use ultrafast lasers in quantum physics research, the datacom and telecom worlds demand maturity, reliability, and low prices. “The idea was always to miniaturize our lasers and fully automate production to address this market,” he says.
Switzerland-based Menhir Photonics is automating production of miniaturized femtosecond laser units, pictured, for the telecom and datacom industry.Menhir
Motivated by growing demand for digital infrastructure, Menhir opened a state-of-the-art production facility in Zurich last August. The facility is developing fully automated assembly machines to produce few-centimeter-square laser modules. They plan to complete their first prototype by the end of 2026, scale up to full automation by the end of 2027, and produce tens of thousands of compact units per year by 2028. “This sovereignty topic is getting important. We realize just how much we rely on China,” Rudin says.
Carlos Lee, director of the European Photonics Industry Consortium (EPIC), is also cautious, pointing out that “sovereignty” is a bit of a smokescreen. Even if Europe does establish semiconductor fabrication plants, it still has to purchase manufacturing equipment elsewhere—and the laser source is just a small part of these. “You’re either 100 percent independent or you’re not. If you don’t have screwdrivers to assemble the machine, you’re not independent,” Lee says. But even if the EU Chips Act and its successor do not lead Europe to more independence, if they help industry, they’re great to have, Lee says: “It’s the market, not sovereignty, that are the real drivers.”
This story was updated on 24 September 2026 to correct the spelling of Litilit.