Australian high-performance computing company Sharon AI announced Thursday that it has secured up to $500 million in funding from blockchain-based financier USD.AI to expand its GPU infrastructure across the Asia-Pacific region.
The funding will support the deployment of computing systems used to train and run large-scale artificial intelligence models, including a $65 million initial deployment expected to begin this quarter, the company said.
According to a press release shared with CoinDesk, the deal will allow Sharon AI to access funding through a non-recourse credit facility, with loans backed by physical GPU hardware rather than the company’s corporate assets.
According to the release, USD.AI’s on-chain lending system turns verified GPU deployments into tokenized collateral, allowing lenders to track performance directly without the need for traditional credit checks.
The structure is designed to allow AI infrastructure providers to grow faster while avoiding slower-growing banks and private market financing, and is a demonstration of how tokenization has empowered private credit markets, the release said.
Maple Finance CEO Sidney Powell told CoinDesk earlier this week that private credit could become a breakthrough use case for tokenization, the process of representing real-world assets on a blockchain.
Private credit markets have limited liquidity and are often less transparent, making them ideal for blockchain-based solutions. Powell said putting real-world loans on the blockchain could improve price discovery and reporting, making the industry more secure and accessible.
Powell, who will be speaking at CoinDesk’s Consensus Hong Kong next month, said he believes defaults in the on-chain credit market will ultimately prove the strength of blockchain rails. He added that transparency around repayments and asset quality could reduce fraud and make it easier for traditional investors to enter the market, especially as crypto-backed loans are starting to be rated by major credit institutions.
USD.AI has currently approved more than $1.2 billion in similar GPU-powered facilities for other AI infrastructure companies, including QumulusAI and Quantum Solutions.
