A year after being forced out of Intel, Pat Gelsinger is still in the middle of the semiconductor wars, waking up at 4 a.m. just because the battlefield has changed. He is currently a general partner at venture firm Playground Global, working with 10 startups. But one portfolio company has really caught his attention. xLight, a semiconductor startup, announced last Monday that it had signed a tentative contract with the U.S. Department of Commerce for up to $150 million, with the government set to become a significant shareholder.
This is a nice feather in the cap for Gelsinger, who spent 35 years at Inter over two stints before the board showed him the door late last year due to lack of confidence in the rebuilding plan. But the xLight deal also puts a spotlight on the Trump administration's trend of acquiring stakes in strategically important companies that has quietly displeased those in Silicon Valley.
“What happened to free enterprise?” California Governor Gavin Newsom asked in a speech this week, capturing the unease that is pervading an industry that has long prided itself on free-market principles.
Speaking at one of TechCrunch's StrictlyVC events at Playground Global, Gelsinger (xLight's executive chairman) didn't seem interested in philosophical discussions. He's more focused on the bet that xLight can solve what he sees as the semiconductor industry's biggest bottleneck: lithography, the process of etching minute patterns onto silicon wafers. The startup is developing a giant “free electron laser” powered by particle accelerators that could revolutionize chip manufacturing. That's when technology works at scale.
“You know, I have a long-term mission to continue to see Moore's Law in the semiconductor industry,” Gelsinger said, referring to the decades-old principle that computing power should double every two years. “We think this is a technology that awakens Moore's Law.”
The xLight deal is the first Chips and Science Act award under President Trump's second term, using funding earmarked for early-stage companies with promising technologies. Notably, the deal is currently at the letter of intent stage, meaning it has not been finalized and details are still subject to change. When asked if the funding could ultimately double the announced amount, or if it might not happen at all, Gelsinger was candid.
“We have agreed on the terms in principle, but as with any contract, there is still work to be done,” he said.
tech crunch event
san francisco
|
October 13-15, 2026
The technology that xLight is pursuing is quite serious in both scale and ambition. The company plans to build a machine approximately 100 meters by 50 meters (about the size of a soccer field) that will be located outside the semiconductor manufacturing facility. These free electron lasers produce extreme ultraviolet light at wavelengths with 2-nanometer precision, much more powerful than the 13.5-nanometer wavelength currently used by ASML, the Dutch giant that completely dominates the EUV lithography market.
Gelsinger explained that in the semiconductor industry as a whole, “about half of the capital is spent on lithography.” “At the heart of a lithography machine is light, (and) this ability to continue to innovate toward shorter wavelengths, higher power light is the essence of being able to continue to innovate toward more advanced semiconductors.
xLight is led by Nicholas Kelez, who has an unusual background in the semiconductor industry. Prior to founding xLight, Mr. Keres led quantum computer development efforts at PsiQuantum (a portfolio company of Playground Global) and spent 20 years building large-scale X-ray science facilities at national laboratories such as SLAC and Lawrence Berkeley, where he served as lead engineer for the linac coherent light source.
Why is this viable now when ASML abandoned a similar approach almost a decade ago? “The difference is that the technology is not yet mature,” Keres explained. At the time, only a few extreme ultraviolet lithography (EUV) machines existed, and the industry had already poured tens of billions of dollars into the current technology. “It wasn't the time to do something completely new and orthogonal.”
The timing seems good, as EUV becomes pervasive in cutting-edge semiconductor manufacturing and existing light source technologies are reaching their limits. The key innovation, Kelez says, is treating light like a utility rather than building it into each machine. “We move away from using tools to build integrated light sources, which is what[ASML]is doing now, which fundamentally limits miniaturization and lower power,” he said. Instead, “we treat light the same way we treat power and HVAC. We build it at utility scale outside the factory and then distribute it internally.”
The company aims to produce the first silicon wafer by 2028 and bring the first commercial system online by 2029.
There are hurdles, of course, but competing directly with ASML doesn't seem like one at this point. “We're currently working closely with them to basically design how we're going to integrate with the ASML scanner,” Kelez says. “So we're working with both them and providers like Zeiss who are responsible for their optics.”
When asked if Intel or other major chipmakers were committed to purchasing xLight's technology, Gelsinger said no. “No one has committed yet, but as you can imagine, work is progressing with everyone on the list and we are in active discussions with everyone.”
On the other hand, the competitive environment is becoming more intense. In October, Peter Thiel-backed semiconductor manufacturing startup Substrate Inc. announced it had raised $100 million to develop a U.S. chip fab that includes EUV tools very similar to xLight's approach. However, Gelsinger doesn't consider them direct competitors. “If Substrate is successful, they could become our customer,” he said, noting that Substrate is ultimately focused on building full-stack lithography scanners that require free-electron lasers, suggesting that is exactly what xLight is developing.
Gelsinger's ties to the Trump administration add another layer to the story. He brought xLight to Secretary of Commerce Howard Lutnick in February, before Playground funded the startup and before Lutnick was granted approval, pitching the company as a company that could help bring chip manufacturing back to the United States.
The arrangement has drawn criticism from some who see the government's approach as state capitalism, where the government picks winners and losers. But Gelsinger is unapologetic, insisting it is necessary for the country's competitiveness. “I measure it by results,” he said. “Does it give us the results we want and do we need to reinvigorate industrial policy? Many of the competitive countries don't have that discussion. They are moving forward with the policies necessary to achieve competitive outcomes.”
He cited energy policy as another example. “How many nuclear reactors are currently being built in the United States? Zero. How many nuclear reactors are currently being built in China? 39. Energy policy in a digital AI economy is equal to a nation’s economic capacity.”
In the case of xLight, government funding comes with minimal conditions. The Department of Commerce has no veto power or a seat on the board, Keres said. “There's no right to information or anything,” Gelsinger added. “This is a minority investment in a non-governmental way, but we can also say that we need this company to succeed in the national interest.”
xLight has raised $40 million from investors including Playground Global and is planning another funding round in January. Unlike fusion and quantum computing startups that require billions of dollars, xLight's path is more manageable, Keres said. “This is not fusion or quantum,” he says. “We don't need billions.”
The company has also signed a letter of intent with the state of New York to manufacture the first machines at its New York CREATE location near Albany, but this agreement also needs to be finalized.
For Gelsinger, xLight is clearly more than just a portfolio company. This is a chance for him to cement his relevance in the semiconductor industry he has built, even if his methods conflict with Silicon Valley's traditional ethos.
Asked how to advance his principles in the current political environment, Mr. Gelsinger retreated to a more technocratic view of corporate leadership: that money comes from the U.S. government, that government is temporary, and that CEOs must stay above the fray.
“CEOs and companies should not be Republicans or Democrats,” he said. “Your job is to achieve your business goals, serve your investors, serve your shareholders. That's your purpose. As a result, you need to understand which policies are beneficial on the R side, which policies are beneficial on the D side, and be able to navigate those.”
He added that apart from the $150 million from the Trump administration, “the taxpayers will do well.”
When asked if working at 10 startups was enough for someone who once ran Intel, Gelsinger emphasized: “Exactly. The idea that I'm able to impact such a wide range of technologies, and I'm a deep tech guy at my core, is really nerve-wracking here. I'm just so grateful that the Playground team is letting me join the team and making them smarter and becoming new venture capitalists.”
He paused for a moment, then added with a laugh: “So I gave my wife back the weekend.”
It's a nice idea, but those aware of Mr. Gelsinger's reputation as a workaholic may wonder how long the arrangement will last.