GPUL — in development.

The planned product for AI Infrastructure Credit: a separate pool of senior secured loans on GPUs and AI servers.

SENIOR SECUREDTITLED EQUIPMENTserials · first chargeASSIGNED REVENUEcontrolled accountIN DEVELOPMENT
Proposed structure

A separate pool, secured on hardware

Separate pool

Own assets, NAV and reporting, fully apart from GBPL.

Senior secured

Loans on GPUs, servers and the revenue they earn.

Record first

External capital follows loans made on our own capital.

How the pool is built

Three phases, each earned by the last

Phase 1 · First loans

Loans made from BridgingFi capital, under the AI infrastructure rulebook.

Phase 2 · Track record

At least six months of repayments at or above 1.25x cover, reported monthly.

Phase 3 · Pool opens

Performing loans pooled into GPUL for professional investors.

How a GPU loan works

From application to repayment

The same discipline as bridge lending: a conservative advance, a defined repayment source and enforceable security.

01

Application

Term sheet, purchase quote, hosting plan and usage evidence.

02

Due diligence

Borrower, demand, data centre, hardware and exit value.

03

Structuring

Security package, controlled account and covenants agreed.

04

Staged funding

Paid to the supplier: deposit, delivery, installation, acceptance.

05

Monitoring

Monthly usage, revenue, uptime and debt-service cover.

06

Repayment

Monthly amortisation from operating cash flow.

Borrower requirements

What a borrower provides

  • Company documents and ownership chart; KYC/KYB on owners of 10% or more
  • Recent financial statements and proof of equity contribution
  • Supplier quotation or purchase order, with specifications
  • Serial numbers for servers and each GPU
  • Colocation / hosting agreement, with lender access rights
  • Customer contracts or usage and revenue history
  • Insurance for the equipment, with the lender as loss payee
  • Installation, burn-in and acceptance test results
Credit controls

How each loan is protected

Conservative advance

Up to 65% of cost at drawdown, stepping down to 55% as the hardware ages.

Amortising

Monthly repayment over up to 36 months, faster than expected value loss.

First charge

Debenture over the equipment and borrower, registered at Companies House.

Controlled revenue

Revenue paid through an account the lender controls; debt service first.

Cover test

Minimum debt-service cover of 1.25x, tested monthly, with a cash sweep below it.

Recovery plan

Insurance, data-centre access rights and a resale route for the hardware.

Pool rules

What a loan needs to enter the pool

Proposed eligibility criteria, applied to every loan before it is pooled.

Proposed pool criteriaPROPOSED
Status
Installed, accepted and earning
Seasoning
At least 3 monthly repayments made
Advance
Current loan-to-value within the step-down schedule
Cover
Debt-service cover of at least 1.25x
Concentration
No borrower above 25% of the pool
Hardware
Current or previous NVIDIA generation
Location
UK or EU data centre
Investor reporting

What GPUL investors would see

01

Monthly

NAV, loan balances, cover ratios, arrears.

02

Utilisation

Usage and revenue per financed server.

03

Collateral

Semi-annual equipment revaluation.

04

Events

Covenant breaches, defaults, recoveries.

GBPL and GPUL

Same platform, separate pools

GBPLGPUL
CollateralFirst-charge UK propertyGPUs and AI servers, with revenue
Loan tenor6–18 monthsUp to 36 months
RepaymentSale or refinanceMonthly amortisation
Main riskProperty value and exitUtilisation and obsolescence
StatusLive · Series 1In development

Follow GPUL

We will share the proposed structure and timetable with eligible investors when it is ready.

Investor access confirmed