USD.AI Handled 20 of 46 Public GPU Financing Deals Worth $5.04B

USD.AI Tallies 46 Public GPU Financing Deals

USD.AI, the on-chain infrastructure provider for GPU-backed lending, says its own aggregation shows it handled 20 of 46 publicly disclosed GPU financing transactions with a combined value of $5.04 billion. The figures come from gpuloans.com, a tracking site the company operates, which counts financings of up to $500 million each announced since January 2023, measured as of November 3. The 46-deal set includes transactions in which GPUs were pledged as collateral as well as financings raised to purchase GPUs where no collateral grant was explicitly stated, so the full count cannot be read as 46 secured loans. Each amount represents the total size of the financing itself, not the exposure actually carried by the participating lenders or by USD.AI.

An interview the company published on November 12 cites a narrower reading: 41 deals and $4.71 billion as of November 2. The two snapshots differ because their cut-off dates and scopes differ, and both should be read as figures from their own aggregation moments rather than as a revision. A single-deal example sits inside the total: the company arranged a $106 million facility to deploy 1,728 GPUs in a Finnish data center, and its largest previously reported GPU financing was a $128.9 million arrangement.

The Midmarket Gap On-Chain Capital Fills

The structure of AI infrastructure finance explains where these deals come from. Large banks handle investment-grade loans in the billions of dollars, while equipment leases below $10 million cover small operators. The gap between the two is the target: GPU cloud businesses that have secured customer usage contracts but are too large for small leases and too small for syndicated bank credit. Within this midmarket, gpuloans.com counts 41 GPU-collateralized financings announced between January 2023 and November 2, 2026, worth $4.71 billion in total with a median deal size of $65 million; USD.AI handled 20 of the 41, about 48.8% of the count, and 32 of the 41 were announced in 2026 alone.

Underwriting runs through Compute IQ, the company's own scoring system, which grades a GPU operator, its customers and its contract terms on a 100-point scale to determine loan eligibility, rate bands, deal size and structure. Liquidity is raised on-chain: depositors place the USDai funding stablecoin and receive sUSDai, whose value tracks the performance of the underlying loan book. The company puts cumulative secondary-market volume across its related tokens at $22 billion as of November 2026, with sUSDai deposits at roughly $550 million in October. Bullish extended a $100 million stablecoin credit line to USD.AI in August and committed to sUSDai trading pairs and a dedicated market-making program. Disclosed risks are falling GPU prices and entry by large banks with cheaper funding costs; USD.AI applies residual-value guarantees through Barkr on selected transactions.

The concentration of deal flow in 2026, 32 of 41 financings inside a single year, indicates a market that is still forming rather than one with a long track record, and COINOTAG's reading is that the deposit structure is what makes the pace possible: with no asset-backed securities market yet built on semiconductors, secondary liquidity of $22 billion against a $4.71 billion origination base is doing the work investors would otherwise exit through. The two risks the company itself names cut in opposite directions. GPU price deflation erodes collateral values that the Barkr guarantees only partially offset, while bank entry would attack the rate advantage the midmarket depends on. sUSDai holders are exposed to both, since the token's value tracks loan performance directly.