They don’t have to do it, but there’s a greater sense of duty at stake. (And sometimes the fate of the universe, too.)

The same thing can happen with founders. Put in the work, find success, cash out, retire to the good life … until something compels them to park the golf cart and get back to tinkering in the garage.

That’s Jason Eichenholz’s story—though in his defense, I’ve never seen evidence of the Brooklyn native tooling around on a golf cart, despite his calling Central Florida home.

More than a decade ago, Eichenholz cofounded Luminar Technologies with Thiel Fellow Austin Russell. After five years under wraps, the automotive technology company—which makes sensors and software that enable self-driving cars—hit the market and booked partnerships with Toyota, then Volvo, then most of the world’s major automakers. The company went public in a $3.4 billion SPAC merger in 2020, making its founders rather wealthy in the process.

Eichenholz eventually relinquished his CTO title and focused on Jonathan’s Landing, his not-for-profit organization committed to building housing for adults with autism. He was focused on philanthropy and legacy building. Life was good. “My mantra since high school was, ‘Why not change the world?’” he says.

Then came the itch—the kind of itch that only a guy with a doctorate in optical sciences and engineering can get.

Eichenholz read about hollow-core fiber technology developed by professor Rodrigo Amezcua Correa at his alma mater, CREOL—the Center for Research and Education in Optics and Lasers within the College of Optics and Photonics at the University of Central Florida. Then he remembered a paper he had written two decades prior about next-generation fiber. “Back then, it was hard to scale,” Eichenholz says. “Frankly, there was no market.”

Amezcua was able to figure out how to send light through air inside a cable in a way that could be manufactured at high volumes—just as the global race for artificial intelligence was heating up. The latest generation of AI demanded expensive chips in large data centers and great amounts of electricity and water. And of course, many miles of fiber cables to connect them.

“This is the first time in my career that I haven’t had to sell the need for my technology,” Eichenholz says. “Either [customers] see the need or we explain the value proposition for the need, and they can no longer unsee it.”

So the pair licensed the intellectual property from the university, booked a multimillion-dollar order from a big customer, hired employees, and ushered a new company into existence called Relativity Networks. “I learned a long time ago not to push technology, but to let the market pull,” Eichenholz says. “I learned a lot with self-driving cars.”

The company exclusively tells Fortune that it has raised $4.6 million in pre-seed funding from private investors, though Eichenholz is quick to add that Relativity is principally relying on the cash flow from purchase orders, rather than copious amounts of outside capital, to grow.

Four months ago, the tech research firm Gartner estimated that power shortages would operationally restrict 40% of existing AI data centers by 2027. With Relativity’s cable, data centers can be located in places they otherwise wouldn’t.

“There’s a rate limiter to the adoption of AI, and that’s power,” Eichenholz says. “Once I can move data down the pipe 50% faster, I can extend the distance between data centers by 50% or extend their geographic optionality,” Eichenholz says. “You can shave years off the opening of a data center site and open it up for green power.”

Relativity is now hunting for strategic partners to make its cable available for hyperscaler customers—think cloud computing providers like Amazon, Google, Meta, et al.—at greater volumes. “I believe we can teach another fiber-optic manufacturer how to make our fiber faster than we can hire those people and build it ourselves,” he says.

In the meantime, the proverbial golf cart continues to gather dust.

“This basically pulled me out of my retirement,” Eichenholz says. “I think the AI revolution will be an order of magnitude more impactful than the internet was, once we figure out the real use cases for AI. I want a front-row seat to what’s gonna happen. I wanna be part of this.”