Former Polychain partners raise $25 million for Ritual, a startup focused on decentralizing AI

And their dominance is something that Niraj Pant and Akilesh Potti, two former partners at the crypto venture capital firm Polychain Capital, are well aware of. “It’s really only a few companies that own the models, that are improving the models,” Pant told Fortune.

He and Potti want to change this, and on Wednesday they announced a $25 million raise for their new startup, Ritual, which hopes to create decentralized computing infrastructure to wrest control of the burgeoning industry away from big tech. Archetype, an early-stage crypto VC, led the raise, and Accomplice and Robot Ventures participated. Pant declined to provide Ritual’s implied valuation.

Pant and Potti’s startup is “approaching real problems at the intersection of AI and crypto, a sector poised for exponential growth in the coming years,” Ash Egan, founder of Archetype, said in a statement. 

Ritual is one of a flock of crypto-AI startups to announce fundraises in the past year, as some Web3 founders have repositioned their companies to ride the current wave of hype for AI and avert the funding downturn for pure crypto firms.

Pant is a tech stereotype—a college-dropout-turned-founder. At 19, he joined Polychain as an intern and dropped out of the University of Illinois one month later. He spent the next six years at the VC outfit, investing in more than 30 different companies, he said, including dYdX and Compound.

Now only 26, Pant decided to team up with Potti, who came to Polychain from Palantir, to launch Ritual.

“Recently, I’ve kind of had an itch to start something and do something in the vein of the types of companies that I would work with,” Pant said, explaining that he had told Olaf Carlson-Wee, the founder of Polychain, that he had aspired to become a founder when he first came on board as a teenager.

At its core, Pant and Potti’s startup is a decentralized cloud computing service for AI. As opposed to companies that exclusively power AI models through computers owned by Google or Amazon, Ritual bands together servers from across many different businesses—even individual laptops, Pant said.

This isn’t a new idea. Others, including Gensyn, Bittensor, Together, and Akash, are also creating decentralized networks for AI. 

However, Pant says Ritual’s offerings differ from other startups’. The cloud-computing network is specifically optimized for inference, or running AI models on a specific input, and fine-tuning, or making small tweaks to AI algorithms. And Ritual also gives companies the ability to prove to customers, through cryptography, that the right model ran. This becomes especially important for showing, for example, that the right medical AI bot ran when providing a cancer diagnosis.

Pant said Ritual has 10 employees, and he’s now looking to add more. “AI talent is quite expensive these days,” he said.