USD.AI Backs Sharon AI With $500M Debt Facility for Compute Growth
What Is USD.AI Financing and Who Is the Borrower?
USD.AI, an onchain lending and stablecoin protocol focused on artificial intelligence infrastructure, has approved a $500 million debt facility for Sharon AI, an Australian provider of AI compute infrastructure. The facility is designed to support the expansion of Sharon AI’s GPU deployments, with the company planning to begin drawing on the credit line this quarter.
The structure reflects USD.AI’s core operating model: lending against tokenized representations of physical GPU assets. Rather than relying on corporate balance sheets or cash-flow history, the protocol accepts verified GPUs as collateral, allowing borrowers to unlock liquidity directly tied to deployed compute infrastructure.
Sharon AI disclosed that it expects to fund $65 million in initial GPU deployments using the facility, targeting expanded capacity to serve hyperscale, research, enterprise, and government customers across Australia and the Asia-Pacific region.
Investor Takeaway
Why GPU-Backed Lending Is Gaining Traction
AI infrastructure remains one of the most capital-intensive segments of the technology sector. High-end GPUs require large upfront investment, while demand cycles are driven by cloud providers, research institutions, and government contracts that often scale quicker than conventional financing can support.
USD.AI positions itself as an onchain alternative to bank lending for this segment. By lending only against verified GPU assets and tokenizing those assets onchain, the protocol links physical infrastructure to programmable liquidity. Borrowers pledge GPUs as collateral, and lenders gain exposure backed by identifiable, revenue-generating hardware rather than abstract credit risk.
According to a Thursday press release, USD.AI has already approved more than $1.2 billion in guidance and non-recourse facilities for AI infrastructure operators, including QumulusAI and Quantum answers. The protocol describes its approach as isolating risk at the infrastructure level, limiting exposure to the specific assets being financed rather than broader corporate liabilities.
This model has attracted attention as AI buildouts accelerate globally and competition for compute capacity intensifies. For operators, it offers access to funding without waiting for sluggisher-moving banking processes. For lenders, it offers collateral that can be monitored and verified onchain.
How the USD.AI Structure Works
At the protocol level, USD.AI operates with a dual-token system built around its USDai stablecoin and a yield-bearing counterpart, sUSDai. Borrowers draw loans denominated in USDai, while can stake into the system to earn yield via sUSDai.
The collateral mechanism is central to the design. GPUs pledged by borrowers are verified and tokenized, creating an onchain representation that ties loan exposure directly to physical hardware. This linkage is intended to reduce amlargeuity around asset backing, a recurring concern in both .
Conor Moore, co-founder of Permian Labs, the firm behind the USD.AI protocol, framed the Sharon AI deal as aligned with the platform’s target user base. “Sharon AI is precisely the type of partner USD.AI was built for – well capitalized with operational expertise and public market discipline, but unwilling to let growth be constrained by sluggisher-moving, legacy financial rails,” Moore said in a statement.
The emphasis on non-recourse structures also limits borrower liability beyond the pledged assets, a feature that may appeal to infrastructure operators managing volatile demand cycles.
Investor Takeaway
What Sharon AI Gains From the Facility
For Sharon AI, the facility offers a way to accelerate deployment without relying solely on equity funding or traditional debt channels. The company confirmed it plans to begin drawing on the loan this quarter, begining with $65 million in GPU deployments.
Sharon AI co-founder James Manning said the partnership fits the firm’s expansion strategy. “We are excited to have partnered with such an innovative financier in USD.AI,” Manning said. “Their approach to GPU leading, and we look forward to further deployments with them as we accelerate our compute infrastructure to service hyperscale, research, enterprise and government customers in Australia and Asia-Pacific.”
The deal also reflects a broader trend in which AI infrastructure providers viewk funding models tailored to hardware-heavy operations rather than software-style growth assumptions. As GPU supply remains tight and deployment timelines shorten, access to flexible financing has become a competitive factor.
What This Means for Onchain Credit Markets
The USD.AI–Sharon AI facility adds to a growing set of experiments linking with real-world infrastructure. Unlike consumer lending or speculative leverage, GPU-backed credit sits closer to project finance, with assets that can be audited, tracked, and valued outside crypto-native markets.
If similar facilities continue to scale, onchain lending protocols may play a larger role in financing sectors where traditional banks move sluggishly or impose restrictive terms. AI infrastructure, with its clear asset base and strong demand signals, is emerging as one of the first testing grounds for that shift.
