Pick up almost any laptop and put your hand near the vent. The warm air pushing out is doing a job first assigned to a spinning blade in the 1800s. Seshu Madhavapeddy noticed that gap between the century and the hardware, and he could not let it go. "The fan was invented in the 19th century," he has said, "and yet as we move into the 21st century, it still remains the main option for heat removal." That single observation became a company.
The company is Frore Systems, and in March 2026 it crossed a $1.64 billion valuation after a $143 million Series D. But the more interesting number is the one he started with: zero moving parts. Frore's flagship chip, AirJet, removes heat using ultrasonic vibration - sound and pressure instead of a rotating fan. It is silent, dustproof, water-resistant, and thin enough to slide into a fanless tablet. Getting there took the kind of stubbornness that most people mistake for foolishness.
01 — The setupTwo decades before the garage
Madhavapeddy did not arrive at cooling as a novice founder. He grew up in the guts of the mobile industry. After a B.Tech from IIT Kharagpur and a Ph.D. in computer science from the University of Texas at Dallas, he ran the smartphones business line at Texas Instruments for five years, then led product and technology at Samsung Mobile, then held a string of VP and general-manager roles at Qualcomm - mobile computing, in-display fingerprint sensors, the Internet of Things.
He had also already done the hard thing twice. His first company, Spatial Wireless, was acquired by Alcatel-Lucent in a deal reported around $300 million. His second, Sipera Systems, was bought by Avaya. By any reasonable measure he had earned the right to coast. He chose the opposite.
Most people, when they heard our idea, would say it's impossible.Seshu Madhavapeddy, on the early days of Frore
02 — The originAn idea that arrived over beer
The specific spark came in San Diego in 2018. Madhavapeddy and Surya Ganti had just finished building a fingerprint sensor together at Qualcomm and were, in his words, "brainstorming about what would be a cool thing to do together if we were to start a company." Ganti had spent years as a research scientist at GE, a company that knows a great deal about moving heat through jet engines. His pitch: borrow the physics of aircraft-engine cooling and shrink it down to the scale of a computer chip.
Madhavapeddy said yes on the spot. They set up in Ganti's garage in Los Altos, named the company Frore - an old English word for frozen - and started trying to build something the industry had quietly assumed was not possible.
03 — The betWhy heat, of all things
The insight underneath Frore is easy to miss because it sounds mundane. A powerful processor throttles when it gets hot - it deliberately slows itself down to avoid cooking. That means the chip you paid full price for often runs below its potential, held back not by silicon but by the inability to carry heat away fast enough. Fix the cooling and you unlock performance that was already there.
For years that was a laptop-and-phone story. Then AI happened, and the same bottleneck showed up at data-center scale, wearing a much bigger price tag. Madhavapeddy had been standing on that exact spot before the wave broke.
Cooling has become the single greatest limiter of AI performance. Traditional thermal technologies cannot keep pace with the AI revolution.Seshu Madhavapeddy, 2026
Frore now talks about a "thermal stack" - pulling heat off the chip and rejecting it into the atmosphere - as foundational infrastructure for AI, the same way people talk about compute and networking. Its newer LiquidJet line pushes into direct-to-chip liquid cooling for GPUs, with claims of markedly higher heat-transfer efficiency and denser racks. AirJet handles the thin-and-silent edge; LiquidJet handles the hot-and-dense core.
Funding rounds toward a cumulative ~$340M. Figures approximate, drawn from public announcements.
04 — The framingFrore's Law
Madhavapeddy borrowed a page from the semiconductor bible and wrote his own commandment. Where Moore's Law promised transistor counts doubling, Frore's Law promises thermal performance doubling every two years. It is part roadmap, part rallying cry, and part warning to the rest of the industry that cooling is not going to sit still.
"We're not bound by the limitations of mechanical devices for how we can make it more and more powerful," he told PC Gamer. It is a quiet kind of confidence - not a promise to beat the fan once, but to keep beating it on a schedule.
05 — The personPatient, technical, allergic to hype
People who describe Madhavapeddy tend to reach for the same qualities: persistence, a long time horizon, and a habit of talking to customers before the product is anywhere near ready. He chased design wins from the first months, when the reasonable reaction to Frore's pitch was polite disbelief. He is more comfortable explaining backpressure and membranes than delivering a keynote flourish.
There is also a genuine cross-disciplinary streak here. It takes a certain kind of mind to sit across a table from a former jet-engine researcher, hear "let's cool a laptop the way we cool a turbine," and treat it as an engineering plan rather than a bar joke. The recognition followed the work: AirJet has been named to the CES Innovation Awards two years running, and Frore has announced collaboration with Intel along the way.
06 — The horizonWhat he is actually chasing
Strip away the funding headlines and the ambition is straightforward: make solid-state cooling the default, everywhere, from a fanless tablet to a rack of GPUs, and keep the performance doubling on schedule so that heat stops being the thing that decides how fast the future can run. It is a big claim from a person who tends to understate things, which is probably why it is worth taking seriously.
Eight years ago the pitch sounded like science fiction and the address was a garage. Today the same idea gets described in press releases as infrastructure for the AI era. Madhavapeddy did not chase that shift. He was already standing where it happened, holding a chip that does not spin.