The race between tech giants to own artificial intelligence is only getting more intense with billions of dollars being spent left and right. So much so, it almost feels like a foregone conclusion that another winner-take-all race is shaping up for one of the established players to own the entire space.

Enter the blockchain projects seeking to ensure that doesn't happen.

As some of those tech giants, like Meta for example, open source their AI models, blockchain projects are popping up to insert themselves as crucial components to hopefully let smaller players compete to win as well. That's the thesis of Allora Labs, which is working on what it calls a decentralized, self-improving machine intelligence network. The project closed its latest round of financing this summer to hit $35 million dollars in total funding.

Coinage recently caught up with Allora Labs co-founder Nick Emmons to explore how crypto, and blockchains inherently plug into the way some of these early AI models are being built.

"I think Bitcoin is essentially commoditized energy," Emmons said. "In doing that, you've created the shortest path possible for someone who has access to excess power or energy and capturing value from it."

Similarly, Emmons explains that Allora is building an onchain process to commoditize an intelligence network for AI models.

"So instead of today's paradigm, where I, as a model creator, need to one build a really good model and then figure out how to bring it to market after productize it, build up business functions, achieve distribution, etc. — Allora offers the shortest path possible from developing a really useful performance model and capturing value from it," Emmons said.

At Meta, Mark Zuckerberg and company have leaned into open sourcing their models and features where others haven't. Some of the ways these models might be able to freely plug and play — or even how these tech companies plan to monetize the tools born out of AI — are still being determined.

As Zuck recently told Nvidia's CEO during a fireside chat, "I don't think that there's just going to be like one AI model, right?" As he predicts, "[Meta] will have the meta AI assistant that you can use, but a lot of our vision is that we want to empower all the people who use our products to basically create agents for themselves."

That could become a major opportunity for Allora and the other projects working to merge value in AI models with a tokenized way to capture that value. If someone were to create something especially valuable, each time it was used by other agents could quickly stack up. Building a network for pieces in that stack, or a platform for people to find them, could also be quite lucrative.

"You're just hooking that model up to this marketplace, instantiated specifically for rewarding useful intelligence, and then capturing value based on how useful your model is," Emmons explained.

But Emmons also shared Zuck's take on avoiding another monopolistic outcome in AI. Elon Musk also falls in that camp, often warning about dystopian outcomes for AI even more loudly than anyone else. For his part, Emmons says democratizing pieces or components of these large models rather than having giant tech companies controlling everything without any outside observation is likely to lead to a better outcome — and one that crypto can power.

"There's a far higher chance of some dire outcome to this if we continue down the path we're on today, where AI is largely seen as a private good that is controlled by a really small number of these centralized monoliths," he said. "I think what crypto offers from a philosophical standpoint and from a practical standpoint, frankly, is a path towards AI that is owned by no one."