Former Intel CEO Pat Gelsinger Bets on Light to Revive Chip Innovation
Pat Gelsinger joins Playground Capital to back semiconductor startups, staking his next chapter on advances in lithography that could push Moore's Law forward again.
Key points
- Pat Gelsinger joined Playground Capital as a general partner in March 2025, after leaving Intel as CEO in late 2024.
- He aims to help semiconductor startups revive Moore's Law, the prediction that transistor counts double roughly every two years.
- Lithography, the process of etching chip patterns with nanometer-scale beams of light, is his chosen battleground.
- Gelsinger took a board seat at xLight, a Playground portfolio company developing new lithography techniques that has received US government investment.
Pat Gelsinger spent 100 days and 100 meetings after leaving Intel before announcing in March 2025 that he'd joined Playground Capital as a general partner. The firm bets on deep tech: startups built on science that hasn't yet been fully proven. He told Wired he was drawn to venture over private equity precisely because the focus stays on the technology rather than the size of the check.
His target is Moore's Law, named for Intel cofounder Gordon Moore, who predicted transistor counts would double roughly every two years. That held for decades, but shrinking transistors to atomic scales has grown prohibitively expensive. Gelsinger's answer is lithography. Today's leading systems, from Dutch company ASML, etch silicon using 13.5-nanometer wavelength light. Free-electron lasers, he argues, could push that down to 5-, 4-, 3-, or even 2-nanometer wavelengths, unlocking far denser chips. "Light is the most important thing," he told Wired.
When Gelsinger joined Playground, he took a board seat at xLight, a portfolio company working on novel lithography techniques that has also drawn US government funding. He's clear the goal isn't to replace ASML: "xLight is not versus ASML, it's with ASML."
What does this mean for the tech industry?
AI has scrambled the semiconductor investment calculus. The industry projected it would reach one trillion dollars in annual revenue by 2030; Gelsinger told Wired he now expects that milestone to arrive next year. We noted in our 15 July story on chip executives and surging demand that infrastructure spending is outrunning supply even as buyers grow pickier about returns. Lithography sits upstream of all of it: every new chip architecture, every memory innovation, every inference accelerator still depends on how precisely light can print a circuit.
That's the real argument here, and it's a patient one. Gelsinger isn't promising a faster GPU by next quarter. He's betting that solving the light problem unlocks everything else: new materials, superconducting components, novel memory stacks. If he's right, the payoff accrues across the whole stack. If the physics proves harder than expected, this is still a long and expensive road. Watch whether xLight's light source can actually integrate with existing ASML machines in a working fab. That's the near-term test that matters.
Common questions
What is lithography in chip manufacturing?
Lithography uses beams of light to etch circuit patterns onto silicon wafers. The shorter the wavelength, the finer the features that can be printed, which means more transistors fit on each chip.
Is xLight trying to replace ASML?
No. Gelsinger has said xLight's first objective is to connect its light source to existing ASML machines to improve them, not compete directly.
