Crypto Billionaire’s Nonprofit Acquires $500M Worth of AI Chips for Generative AI

Crypto billionaire Jed McCaleb, known as a co-founder major crypto companies like Stellar, Ripple, and Mt Gox, has launched a nonprofit with a mission to empower the development of generative AI.

This nonprofit, registered in the US as a 501(c)(3) organization, recently purchased 24,000 Nvidia H100 graphics cards (GPUs), estimated to be worth a staggering half a billion dollars, TechCrunch reported on Tuesday.

The powerful GPUs, among the largest in the world, will fuel data centers that are set to lease computing capacity to AI projects.

These data centers are not the usual technology giants that typically provide computing power for AI development; instead, they represent a new approach led by McCaleb, the report said.

Startups such as Imbue and Character.ai are reportedly already leveraging the computing capacity for AI model experimentation.

According to Eric Park, CEO of Voltage Park, the organization responsible for managing these data centers, the goal is to “unlock access” to AI resources, enabling startups, scale-ups, and research organizations to innovate without the constraints of restrictive contracts, GPU scarcity, and high minimum purchase thresholds.

“We’re continuing to speak with folks in the industry to understand their needs and are using what we learn to inform how we build out our remaining clusters so that they’re useful to as many customers as possible,” Park told TechCrunch.

Ownership transferred to Voltage Park

McCaleb’s nonprofit, the Navigation Fund, reportedly purchased the GPUs and then transferred ownership to for-profit Voltage Park, which is a subsidiary of the Navigation Fund.

While this structure may appear complex, it allows for-profit Voltage Park to efficiently manage the large GPU cluster, TechCrunch explained.

While Voltage Park’s full cluster is yet to go online, it will be accessible across multiple US states, offering both short-term leases and hourly billing.

GPU shortage a significant challenge

The shortage of advanced compute capacity, especially GPUs, is currently a significant challenge in the AI industry.

Among the companies struggling to secure these chips is Microsoft, which has warned of service disruptions due to a shortage of hardware for AI operations.

According to the report, Nvidia’s high-performance AI cards are sold out until 2024.