AI is global. Your capital should be too.
Institutional financing for contracted compute, typically $50 million to $2 billion and above per transaction, with capital drawn from markets around the world.
Two ways in
For operators
Neoclouds, inference providers and data center operators deploying GPUs against a contract with a named customer.
Operator financingFor lenders
GPU lenders, equipment finance firms and specialty finance platforms. We structure institutional refinancing of the senior portion of your loans, so you can redeploy the cash.
Lender programsWhere the gap is
GPU financing is served at the two ends of the market: corporate balance sheets for the largest buyers and equipment leases for small clusters. Between them sits a large and growing market of mid-sized deployments with few institutional lenders.
Our notes are built for institutional credit investors in the United States and Europe, including insurance-owned credit platforms, asset managers and credit funds, that buy senior secured paper backed by contracted cash flows.
GPU financing for operators
How the GPUs are funded, illustrative
For the operator
Senior secured financing for GPUs, networking and installation. We structure financing for the purchase against the contract, so the hardware can be ordered at signing rather than after a separate equity raise. Repayment comes from the contract’s monthly payments, and the security is the equipment and the contract itself.
For the customer
Many compute contracts require the customer to prepay 15% to 25% of the contract value at signing, often from its most expensive capital. We can structure financing for part of that prepayment for the customer. The operator still receives the full prepayment, and the customer repays the financing as the prepayment is credited against its monthly bills.
Prepayment financing, an illustrative example
How the money moves
At signing
Over the contract term
Illustrative. Structures depend on each contract and on the institutional capital behind it.
Before you sign
Many compute contracts are negotiated without the terms institutional lenders require, and those terms are hard to add after signature. These are the ones that decide whether a contract can be financed.
At least as long as the financing.
Fixed payment for reserved capacity, independent of usage.
Customer consent to assignment as security and to payment into a lender-controlled account.
Sized to cover the outstanding financing on early termination or customer default.
Downtime credits capped below the level that impairs debt service.
Parent guarantee, letter of credit or larger prepayment where the customer is not investment grade.
Lender or replacement operator step-in at the site.
A financing readiness review: we read the draft contract before signature, mark the terms the institutional capital will need, and propose language for each, so the financing is ready when the contract is signed.
Compute loan refinancing for lenders
How a senior refinancing works, illustrative
We structure institutional refinancing of the senior portion of compute loans you already hold, as an A-note participation or the senior tranche of a pool, typically $50 million to $2 billion and above per transaction. You receive cash to lend again, and you keep the borrower relationship and the servicing.
Lenders originating compute loans often reach the limit of their own balance sheet before they reach the limit of their pipeline. Refinancing the senior portion turns a full book into capacity for new loans, without selling the relationship.
For larger loans, the senior A-note is placed with institutional capital under a co-lender agreement. You keep the servicing.
Loans are pooled and the senior tranche is placed, with the servicing staying with you. A pool can be sized to grow with origination.
Where value insurance from an investment-grade carrier covers the balance at maturity, the insured exposure is structured as senior.
Loan tape (advance rate, amortization profile and any balloon, tenor, offtaker profile, reserves, payment history), loan and security documents, servicing and collateral monitoring arrangements, and any value insurance policy.
How it works
Describe the contract or the loan, the deployment and the capital required through the confidential form below.
We review every submission against the criteria of the institutional capital we work with and reply directly.
Where the contract supports it, we design the financing around the transaction and bring it to institutional capital.
At closing, the capital funds the GPU order and, where agreed, the customer prepayment. Repayment runs from the contract cash flow once the deployment is live.
Guides
How a signed compute contract can repay the debt that funds the GPUs.
The contract clauses that decide how much a lender will advance.
Why lenders treat reserved and on-demand revenue differently, and how financing scales.
Financing part of the customer’s prepayment while the operator receives it in full.
Separating what the GPUs are worth at the end of the term from the senior debt.
Financing the building, the power and the GPUs each on their own terms.
Common questions
A capital markets firm whose team brings together decades of experience across the disciplines that compute and data center financing draws on: asset-backed structuring, energy and power infrastructure, real estate development, and institutional capital. Between them, the team has worked at Merrill Lynch, Morgan Stanley, Bridge Investment Group, Valero Energy and Google, served as a U.S. Navy nuclear submarine officer, and led more than fifty large-scale real estate projects across the United States, Europe and the Caribbean. The firm is based in Spain and the United States. Meet the team.
No. We structure the financing and arrange it with institutional investors, which hold the notes. That lets the size of a financing follow the transaction rather than our balance sheet.
Because the contract is the source of repayment. Institutional senior investors will lend against contracted payments. Most will not lend against what the GPUs might be worth, or might earn, after the contract ends. Where the hardware’s value at maturity is insured, a balance at maturity can be considered.
The lower of two measures: a share of the equipment cost, and the amount the contract’s cash flow can repay within its term with room to spare. For a three-year contract, the second measure usually decides. Longer contracts, larger prepayments, stronger customers and value insurance all increase the amount.
Yes, with the customer’s total compute commitments across providers disclosed. These contracts usually need more credit support, a larger prepayment, or a pool that spreads the exposure across several customers.
Security over the equipment and the assigned contract, a funded debt service reserve, offtake interruption and value insurance where available, and step-in rights protect the debt while the capacity is re-let to another customer.
The United States, the United Kingdom and Europe, with NVIDIA and AMD hardware.
Yes. Submissions are received in confidence, and customer names may be withheld until a confidentiality agreement is in place.
Submit an opportunity
Submissions are received in confidence. Customer names may be withheld at this stage.