AI infrastructure is paid for by a mix of the largest technology companies’ own spending and outside capital: bond and loan markets, private and project finance, equipment leases, customer prepayments, and, on the power side, utilities that build generation and storage under long-term contracts with data-centre customers. The International Energy Agency (IEA) says the scale now exceeds what company balance sheets can fund alone. The mix cannot yet be measured as a single split, and the final section explains which parts of the picture current evidence leaves open.
Big technology companies: the starting point
The clearest hard figure is the IEA’s count of capital expenditure by five large technology companies. It exceeded USD 400 billion in 2025, and the agency expects it to rise by a further 75% in 2026 (International Energy Agency, 2026). Two qualifications matter. The figure covers five selected companies, not all AI investment worldwide. The 2026 number is an expectation, not a reported total. The figure also does not show how each company paid for that spending, which is the subject of the sections that follow.
As an Amazon Associate I earn from qualifying purchases.
Why company balance sheets are no longer enough
The IEA’s assessment is direct:
“Data centre investments have grown too large to be funded from company balance sheets alone, and large amounts of funding from capital markets will be critical for their buildout.”
Free tools Windows power users keep installed
One-click scans. No signup required.
Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.
Source: International Energy Agency, Key Questions on Energy and AI (2026).
#1 Best Overall
That shift changes the question. Once capital markets carry a significant share of the buildout, its pace depends on investor appetite as well as on the cash the largest firms generate.
Debt: an estimate of need, and a warning
The Bank of England’s July 2026 Financial Stability Report says AI companies are increasingly turning to external finance, particularly debt, to fund infrastructure. It relays a Barclays estimate that USD 240 billion of AI hyperscalers’ 2026 investment needs would be financed through investment-grade credit issuance. That is an attributed estimate, not a final observed total of bond issuance.
The Bank’s concern is the risk side. It warns that debt servicing and opaque financing structures could create financial-stability risks. Those risks are tied to whether the revenue behind the investment arrives as expected, a point the IEA also makes about returns.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Infrastructure providers: GPU financing and customer prepayments
Below the largest cloud companies sit providers that build and rent compute capacity. IREN’s FY2026 SEC filing is one example. It says its customers include hyperscalers, frontier labs, AI developers and enterprises, and it describes GPU financing and customer prepayments as sources supporting deployment.
Rank #2
One filing shows what a single provider uses, not how the market as a whole is financed. Prepayment is the feature worth noting. When a customer pays ahead of delivery, part of the funding burden sits on the customer side of the contract rather than on the provider’s own cash or on bonds.
Leases, project debt and special-purpose vehicles
A 2026 NBER working paper lists leases, joint ventures, project debt, private credit, securitization and special-purpose vehicles (SPVs) as financing channels worth examining. Its buildout estimates are model-based projections, and its list is an analytical framework, not a ranking of how much each channel is used. Its practical lesson is structural: the same data centre can be financed by a corporate owner, a project company or a lessor, and each choice places ownership and risk differently.
Power: utilities and who signs for the load
Electricity is where the financing question becomes a contract question. AEP’s 2026 investor presentation describes long-term agreements with data-centre customers that include minimum monthly charges and customer credit and collateral protections. Its plans include substantial storage and generation investment to accommodate expected loads.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesThese provisions show how a contract can assign risk to the customer. A minimum charge is owed regardless of how much power is drawn, and collateral backs the commitment. They do not establish who ultimately bears every system-wide cost of new capacity. That depends on tariffs, the costs that pass to other ratepayers and contract terms that a single presentation does not describe.
Timing adds pressure. The IEA’s 2025 analysis notes that energy-system infrastructure has longer lead times than data-centre construction, so power capacity may need to be financed before the load it serves is built.
What the money buys
The IEA’s 2025 analysis lists the main components of data-centre infrastructure:
- Servers
- Networking equipment
- Cooling systems
- UPS batteries
- Backup generators
- Grid connections
These parts differ in useful life and lead time. Financing for compute hardware is therefore not the same problem as financing for the grid connection that serves the building.
Demand and returns: why the pace depends on expectations
The IEA reports that data-centre electricity demand grew 17% in 2025, while electricity demand at AI-focused data centres grew 50%. Its updated projection puts data-centre electricity use at 485 TWh in 2025 and 950 TWh in 2030. These are the agency’s estimates and projections, not financing totals, but they explain why power is part of the investment question.
The IEA also ties the pace of build-out to sentiment:
“The pace of data centre growth, and the resulting increase in energy consumption, will be sensitive to market sentiment, including expectations for returns on investment in data centres and AI deployment, as well as to broader macroeconomic and financing conditions.”
Source: International Energy Agency, Key Questions on Energy and AI (2026).
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →That dependency is the central uncertainty. The evidence identifies it but does not show whether current spending will earn adequate returns, or that it will fail. The IEA and the Bank of England both treat return expectations and refinancing conditions as live risks.
Best Value
Comparing the channels
The table sets each channel beside the evidence attached to it. “Not stated” means the cited source does not give that value. It does not mean the value is zero or small.
| Channel | Evidence in the cited sources | What the sources do not establish |
|---|---|---|
| Corporate cash flow and capital expenditure | IEA (2026): capex by five large technology companies above USD 400 billion in 2025, with a 75% increase expected in 2026 | Share of total AI investment; how each company funded its capex (not stated) |
| Investment-grade bonds | Bank of England, July 2026 Financial Stability Report, citing a Barclays estimate of USD 240 billion of 2026 hyperscaler investment needs to be financed through investment-grade credit issuance | Whether the estimate matches actual issuance (not stated; the figure is attributed and is an estimate) |
| GPU financing and customer prepayments | IREN FY2026 SEC filing, describing both as sources supporting deployment | Market-wide scale; terms used by other providers (not stated) |
| Leases, project debt, private credit, securitization and SPVs | NBER working paper (2026), as an analytical framework | Relative size of each channel; buildout figures are model-based projections (not stated) |
| Utility investment under long-term contracts | AEP investor presentation (2026): minimum monthly charges, credit and collateral protections, storage and generation plans | Who ultimately bears system-wide costs; one company example does not settle this |
How to read a financing deal
When a company or utility discloses an AI-related financing arrangement, these points determine who pays:
- Who owns the data centre, the compute equipment and the power asset.
- Whether the customer prepays, and on what terms the prepayment applies.
- Minimum charges, contract duration and termination protection.
- The collateral or credit support the customer must post.
- Who funds generation, storage and connection capacity, and which costs pass to other ratepayers.
- Who carries utilization, refinancing and residual-value risk if demand, technology or capital-market conditions change.
What cannot be measured yet
Three answers readers often want are not available from current public evidence:
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minute- A global split of AI infrastructure investment among corporate cash, bonds, private credit, customer prepayments and utility capital. The sources do not give a consistent one.
- How much current investment is already covered by recurring customer revenue.
- A single total built from company capex forecasts, single-issuer disclosures and model projections. These differ in scope and date, so adding them would produce a figure no source states.
When a headline share appears, check its scope, its date, and whether it measures spending, commitments or bond issuance.
Quick Recap
What to watch
- Updates to the IEA’s capex and electricity projections, which will show whether the expected 2026 increase materialises.
- Future Bank of England Financial Stability Reports, for changes in the debt-servicing and structural risks it flags.
- Company filings that disclose prepayments, minimum charges and collateral, which are the most direct evidence of who carries risk in a given deal.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
