BroadChain News, April 24, 02:16, As the AI industry expands rapidly, the entire infrastructure sector, from NVIDIA GPUs to data centers, is experiencing an unprecedented capital expenditure cycle. However, assets like GPUs have an inherent contradiction: they generate cash flow quickly but also depreciate rapidly. A popular GPU may become significantly obsolete within three years, and traditional banking systems tend to be cautious about assets with high depreciation, high volatility, and rapid technological iteration.
Against this backdrop, a wave of projects attempting to bring AI infrastructure financing onto blockchain has emerged. USD.AI is a relatively unique one—it does not build models or AI agents but instead creates an on-chain credit protocol centered around GPU financing. Its governance token, CHIP, has received support from major exchanges including Binance, Bybit, Coinbase, and Upbit. According to CoinGecko data, as of today, CHIP's circulating market cap is approximately $215 million, with an FDV of $1.08 billion, and it has risen about 57.7% since its launch.
USD.AI is developed by Permian Labs, with the USD.AI Foundation providing off-chain governance and legal infrastructure. Its core logic can be summarized as "off-chain organization, on-chain lending," channeling on-chain dollar liquidity into capital-intensive expenditures such as GPUs, server racks, and data center expansion. Borrowers must first place purchase orders with OEMs and establish SPVs. After GPUs are delivered and installed in data centers, electronic receipts are generated, which are tokenized into ERC-721 NFTs and ultimately used as collateral to obtain USDC loans within the protocol.
Publicly verifiable cases are still limited, but some hardware-related loans have been disclosed. In April 2026, Crucible Capital withdrew a $26.82 million loan, collateralized by 72 units (576 GPUs total) of NVIDIA B300 GPUs, deployed in Washington State, USA. However, the project has not fully disclosed the list of OEMs, data center operators, or borrowers. Currently identifiable entities include QumulusAI, Quantum Solutions, Sharon AI, and Crucible, which are not consumer-grade AI products but computing infrastructure providers.
USD.AI's innovative GPU stablecoin lending model has also garnered institutional support. In August last year, USD.AI announced the completion of a $13 million Series A funding round, led by Framework Ventures, with participation from Bullish, Dragonfly, Arbitrum, and others. However, it is worth noting that while the official team repeatedly emphasizes a focus on reviewing the quality of offtake agreements, it has not yet disclosed customer lists, contract numbers, or actual leasing scales—making it difficult for outsiders to independently verify whether these GPUs can generate stable and sustainable cash flows.
USD.AI's structure is divided into three layers: USDai is the liquidity layer, sUSDai is the yield layer, and CHIP is the governance layer. USDai, as the "base layer" for on-chain financing, is pegged 1:1 to the US dollar, designed for maximum compatibility with DeFi protocols and does not generate yield. Users can deposit PYUSD to mine USDai and redeem it at any time. sUSDai, on the other hand, corresponds to long-term GPU loan positions, involving asynchronous redemptions, exit queues, and liquidity restrictions, with risks entirely different from USDai. CHIP holders gain voting rights by staking sUSDai to participate in protocol governance.