Sharon AI signs a $356 million GPU-backed debt facility at 9.95%
Source report: 2026-10-01 · Editorial analysis published: 2026-10-02
Two Australian subsidiaries can borrow through secured tranches totaling $356 million to fund contracted compute deployments; the fixed rate excludes fees and the full amount is capacity, not cash already spent.
Analysis and practical implications
This section is our analysis and illustrative calculations, separate from the source report.
The agreement creates two secured loan tranches
Sharon AI announced on October 1 that two Australian subsidiaries entered a senior secured term-loan facility totaling up to $356 million. The SEC filing divides it into Facility A of up to $150 million and Facility B of up to $206 million. Up to is important: it describes borrowing capacity under the agreement, not proof that all funds were drawn on the signing date or immediately converted into operating GPU capacity.
The headline fixed rate excludes fees
The company states a fixed interest rate of 9.95%, excluding fees. If the entire $356 million were outstanding for a full year, simple interest at that rate would be about $35.4 million before arrangement fees and other costs. That arithmetic is illustrative, not company guidance, because actual interest depends on draw timing, principal outstanding, repayment and the detailed facility terms. Operators comparing financing should use the all-in cost, not the headline coupon alone.
GPUs and associated cash flows support the debt
The facility is described as GPU-backed and secured against equipment and associated cash flows. The SEC filing says proceeds can finance or refinance servers, CPUs, GPUs, networking, storage, related infrastructure, debt-service reserves and transaction costs at contracted data-center facilities. This links borrowing to identifiable assets and customer revenue, but collateral does not eliminate technology depreciation, customer-credit, utilization, delivery or execution risk.
The borrowers and guarantees are ring-fenced
SAI AU No. 1 and SAI AU No. 3 are the borrowers. Their immediate holding companies guarantee the obligations, while Sharon AI provides a limited and releasable guarantee under the filed structure. Using special-purpose borrowers can isolate project assets and cash flows, but readers should not automatically call the debt non-recourse without following the precise guarantees, security package, covenants and conditions in the definitive documents.
The money is intended for contracted compute deployments
Sharon AI says proceeds will support compute infrastructure dedicated to customer contracts. Contract-backed financing can align debt service with predictable payments, but total contract value is not the same as recognized revenue or free cash flow. Deployment still requires hardware delivery, data-center readiness, acceptance tests and customer performance. The company’s statement does not turn future contracted value into revenue on October 1.
The company names institutional capital providers
The release says the facility includes Australian, Asian and global investors, including Goldman Sachs and selected private-credit funds, and names Jarden Australia as sole financial adviser and arranger. The announcement does not disclose every lender’s commitment, draw schedule or fee. The presence of well-known institutions is relevant to financing access, while the filed agreement and later financial statements remain the better evidence for amounts actually borrowed and interest paid.
A larger GPU target remains forward-looking
Sharon AI calls this the first in an expected series of GPU financings supporting a scheduled buildout of more than 68,000 NVIDIA GPUs by mid-2027. It also says it secured more than $2.6 billion of institutional debt and equity capital over ten months. These are company-reported financing and deployment statements. The 68,000-unit figure is a target, not an installed-fleet count, and should be tracked against deliveries, energization and customer acceptance.
The structure matters to the wider compute market
Bitcoin miners and AI operators compete for powered sites, transformers, cooling equipment and capital. GPU-backed lending shows how AI infrastructure can finance hardware through customer-linked cash flows instead of relying only on corporate equity. ASIC fleets can also be financed against equipment or production, but their residual values and volatile mining revenue differ from contracted AI services. The comparison should focus on collateral quality, cash-flow certainty, useful life and refinancing risk.
What is confirmed and what to monitor
Confirmed facts are the signed facility, its two tranches, maximum $356 million size, stated 9.95% fixed rate excluding fees and asset-and-cash-flow security. The release does not prove full drawdown, completed GPU installation or future profitability. The next checkpoints are borrowings reported in financial statements, interest and fee expense, GPUs delivered, megawatts energized, contracted capacity accepted and revenue recognized. Those measures will show whether financing capacity becomes productive compute.
Source: Sharon AI ↗
Mining calculator ↗
