Institutionally oriented digital assets platform Bullish (NYSE: BLSH) has committed a $100 million stablecoin credit line to USD.AI, a protocol that originates loans against high-performance computing hardware.
The facility is intended to fund non-recourse financing for operators building artificial-intelligence infrastructure, with repayment claims anchored to graphics processing units rather than a borrower’s entire corporate balance sheet.
The arrangement sits at the junction of two expanding markets: private credit for the enormous capital spending required by AI data centers, and the tokenization of physical assets so that crypto liquidity can reach those projects.
GPUs have become some of the most valuable pieces of equipment in the technology sector because they handle the parallel processing that large models demand.
By treating clusters of those chips as identifiable collateral, USD.AI aims to give infrastructure companies access to capital without forcing them to dilute equity or pledge unrelated assets.
Bullish, which is listed on the New York Stock Exchange and owns CoinDesk, described the AI compute financing market as already larger than several familiar consumer-credit categories.
Thomas Cowan, the firm’s head of tokenization, said the decision reflected a longer-standing view that well-structured real-world assets belong on public blockchains.
On-chain transparency, he added, allowed Bullish to underwrite the facility with the same diligence it applies elsewhere on its platform.
USD.AI, developed by Permian Labs, structures loans so that a bankruptcy-remote vehicle typically holds the financed servers and the protocol takes a first-priority security interest in the hardware.
Transactions settle on-chain, giving lenders a clearer line of sight into the collateral.
The protocol already had more than $225 million in crypto assets locked at the time of the announcement and had recently completed sizable facilities backed by current-generation Nvidia chips.
As part of the partnership, Bullish Exchange intends to list USD.AI’s yield-bearing token sUSDai on multiple trading pairs and support those markets with a dedicated market-making program.
The goal is to create secondary liquidity and more transparent price discovery around the cost of compute-backed credit.
The two organizations also plan to expand joint research on capital-formation models for AI capital expenditure.
The model is not without risk. GPU prices can move sharply, and successive generations of chips can reduce the residual value of older inventory.
Utilization rates and customer contracts therefore matter as much as the hardware itself.
Coverage of the deal has noted that the real test will come if a borrower defaults and lenders must rely on the physical collateral.
Even so, the transaction illustrates how regulated crypto market infrastructure is being used to channel institutional capital into one of the most capital-intensive build-outs of the decade.
For operators, the structure offers non-dilutive funding tied to income-producing machines. For lenders, it offers direct exposure to those machines rather than to the broader credit of a young infrastructure company.