How Akash and Bittensor Rode the AI Wave to Become Crypto's DePIN Darlings

This year, Coinage is highlighting the best projects in crypto as part of our annual Crypto Project of the Year celebration. Our members will vote on which one of the 16 nominated projects will win the crown. Head to Coinage's Snapshot to vote — ( you'll need a Coinage Caucus or Network NFT.)

Up first, Akash and Bittensor:

• Akash, now a two-time Crypto Project of the Year nominee, has seen its native AKT token surge more than 130% since last year as the network grows its two-sided decentralized computing marketplace — and sees a boom from AI training use cases.

• Bittensor, a first-time nominee, has seen its TAO token double over the last year as it seeks to pioneer the decentralized production of artificial intelligence.

Akash co-founder Greg Osuri joined Coinage to discuss what has been fueling a more than 1,000% rise in network fees, and how he sees the AI explosion continuing to fuel a need for more computing power. As large language models like ChatGPT and Claude continue to proliferate, Akash has positioned itself as a critical player in addressing shortages and inefficiencies.

While NVIDIA has scaled up production of its coveted H100 GPUs, Osuri notes the market remains fragmented. Akash’s decentralized marketplace provides a solution, aggregating smaller pools of compute resources into a cohesive network that rivals traditional data centers.

"So last time we were here, I remember we talked about GPU shortage, right? There was a clear shortage for GPUs ... and Nvidia since has scaled up supply quite a lot ... so the supply chain problem has eased quite a bit compared to last year, but it created a new problem which is getting the right supply at the right place," he said. "So that gives an opportunity for marketplaces like Akash to be able to be the one stop shop to get access to these hundreds of data centers that have small quantities."

But hardware isn’t the only bottleneck. Energy requirements for training large language models (LLMs) have surged, with next-generation models demanding gigawatt-scale capacities that existing infrastructure struggles to meet. It also requires massive data sets — data that could potentially be more easily sourced and formatted.

Bittensor, a decentralized, open-source, and permissionless machine learning network, is looking to help in that regard. The project claims it is transforming machine intelligence into a tradable commodity. And recently, Bittensor also won big support from Barry Silbert — the disgraced billionaire founder of Digital Currency Group and former Chairman of Grayscale. Silbert is now running Yuma, a DCG subsidiary that is looking to compete with Google and OpenAI in the field of artificial intelligence:

Speaking with Fortune, Silbert likened decentralized AI to the world wide web, which in the 1990s supplanted the "walled garden" version of the Internet run by a handful of tech firms. It's unclear if a decentralized model of AI can hold its own in an industry where the leading firms depend on massive amounts of data, high priced chips and computing power. But Silbert says he is convinced a permissionless version of AI is better—so much so that he is becoming a hands on CEO for the first time in four years to lead Yuma.

However, if there's one area both projects appear to align, it's a commitment to being fully open source. “Open source attracts the best minds,” Osuri explained, noting that over 500 contributors from institutions like Google and Microsoft are actively shaping Akash’s development. This decentralized ethos extends to security, where transparency ensures rapid identification and resolution of issues —a stark contrast to the vulnerabilities of closed systems.

Decentralization is not just a principle but a practical advantage. Even if core contributors step away, Akash’s community-driven model ensures the network’s resilience. “Our goal is to make ourselves irrelevant,” Osuri remarked, reflecting the project’s vision of long-term sustainability.

Now, it may be difficult to distinguish which project here could be a purer play on AI, or whether TAO and Akash actually compliment eachother (which they likely will.) But in the shorter-term, or at least in weighing which project should advance in our Crypto Project of the Year celebration, we turn to you — our community.

Coinage NFT holders can vote now in our Snapshot vote to advance Akash or Bittensor to the next round. Voting closes December 9, mint a Coinage NFT now to have your voice heard!