The push for public AI computing power.

From American Public Media, this is Marketplace Tech.

I'm Nufa Sapo.

If you want to power AI, you need graphics processing units or GPUs.

The richest tech companies are investing billions of dollars to amass thousands,

sometimes millions of these chips.

But for those without deep pockets, that kind of computing power is out of reach.

That's why the state of California recently launched an initiative called

Cal Compute to look into building its own public GPU cluster for startups and nonprofits.

There are similar pilots in New York State and at the federal level. Marketplaces Megan

McCarty Carrino has more.

The Computer History Museum in Silicon Valley takes you through 2,000 years of computing history,

from the abacus to the iPhone. Curator Mark Weber gives me a tour of the machines that remade

modern life. This is mainframes, the IBM 360. This little spool of paper tape with holes in it is the origin

of Microsoft. He says the history of computing isn't just the story of the scrappy startups

that became household names, but of cross-pollination between industry, government, and universities.

In my mind, I see them all as complimentary.

It's kind of an ecosystem where you need all the parts to make it work.

But lately, that diverse ecosystem has been breaking down.

Only a few private companies can now afford to develop and study the biggest, most advanced, artificial intelligence models.

There's a lot of raw potential that sits out there right now to have incredible scientific breakthroughs if we were able to resource appropriately.

Russell Wald is the executive director of the Stanford Institute for Human-Centered AI.

...