🔍 Read the full analysis: From Cash To Power Bills: Who Pays For AI Data Centers? on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Amazon, Microsoft, Alphabet and Meta spent about $202 billion on property and equipment in July-September 2026, more than twice the amount a year earlier, according to the source’s calculation from company filings. Their funding mix now includes operating cash, debt, leases, partnerships and equity, while power contracts and other long-term commitments can shift risk to investors, landlords and electricity customers.
Amazon, Microsoft, Alphabet and Meta spent about $202 billion on property and equipment in the third quarter of 2026, more than twice their spending a year earlier, according to a calculation based on company cash flow statements. The scale of the AI infrastructure build-out is pushing the companies to combine cash with borrowing, leases, partnerships and equity financing, spreading some of the financial exposure beyond their own balance sheets.
The source’s calculation puts the four companies’ combined third-quarter spending at about $202 billion, compared with roughly $97 billion in the same period of 2025. Amazon reported $77.3 billion in property and equipment spending, Alphabet $51.4 billion, Microsoft $39.4 billion in cash spending and Meta $33.9 billion. Microsoft said its total capital spending, including new finance leases, was $45.2 billion.
The companies have also disclosed new financing. Meta said it sold $25 billion of stock and $12.5 billion of preferred shares during the week covered by the source. Microsoft issued $25 billion in bonds, including debt with a maturity in 2066. Amazon disclosed a loan facility of up to $80 billion secured by servers it plans to buy. These arrangements differ in cost and structure, but each helps fund infrastructure without relying solely on current operating cash.
Future commitments add another layer. The source’s sum of four quarterly filings puts leases that had been signed but had not yet started at about $1.02 trillion as of September 30, 2026, mostly for data centers. That figure is a total of future payments, not a bill due immediately. Meta said on October 26 that it planned to fund AI investment through “a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.”
Infographic 01 · One quarter, four companies
The bill doubled. The cash did not.
What Amazon, Microsoft, Alphabet and Meta paid for buildings, servers and chips from July to September 2026, set against the cash their businesses brought in.
AmazonQ3 2025: $35.1B spent
Spent $34.9B more than came in
AlphabetQ3 2025: $24.0B spent
Free cash flow −$9.1B
MicrosoftQ3 2025: $19.4B spent
$18.6B left over
MetaQ3 2025: $18.8B spent
Free cash flow −$2.1B
Cash purchases of property and equipment as reported; not all of it is for AI (Amazon also builds warehouses). Gaps for Amazon and Microsoft are my arithmetic; free cash flow for Alphabet and Meta as the companies report it. Microsoft’s quarter is its fiscal Q1 2027 (July–September 2026).
Sources: Amazon results, 29 Oct · Microsoft 10-Q, 28 Oct · Alphabet results, 28 Oct · Meta results, 21 Oct. Debt raised: cash flow statements, nine months 2026.
Powered by Thorsten Meyer AI
How Infrastructure Costs Spread
The financing mix matters because a data center’s costs can outlast the equipment it houses. The source says servers are generally written off over five to six years, while power contracts may last 15 to 20 years, data center leases can extend up to 20 years, and some bonds run for 40 years. If demand or AI revenue falls short of expectations, companies and other parties may still owe payments tied to assets that are no longer earning as much.
The immediate exposure is not evenly distributed. The four large technology companies have substantial businesses and access to capital markets, but financing can also involve landlords, infrastructure companies, lenders and investors. The source points to GPU cloud provider CoreWeave as an example of a newer infrastructure company with senior notes carrying interest rates of 8.5% to 9.75%. Those rates are reported for that company’s notes, not a general borrowing rate for the sector.
Electricity users may also be affected, though the source does not quantify any resulting household bill increases. Long-term power contracts and demand for new electricity capacity can connect data center expansion to utility planning and local energy costs. Whether those costs are borne by the technology companies, project partners or other customers depends on contract terms and local rules.
Products Worth Considering
Jackery Explorer 300 Portable Power Station, 292Wh
Jackery Explorer 1000 v2 Portable Power Station
Anker SOLIX S2000 Portable Power Station, 2,010Wh, 1,500W
As an affiliate, we earn on qualifying purchases.
The Spending Behind the Financing
The source describes a change from earlier years, when large technology companies could largely fund data centers from profits generated by advertising, software and online shopping. In 2026, capital spending expanded sharply as companies built out facilities, bought servers and chips, and prepared for AI services. The reported totals cover property and equipment, which can include assets beyond data centers; Amazon, for example, also invests in warehouses. The company said the increase primarily reflected AI investment.
Company guidance points to continued spending, though forecasts are not guarantees. Amazon expected about $220 billion in cash capital spending for 2026. Alphabet maintained a range of $195 billion to $205 billion and forecast a “significant” increase in 2027. Microsoft expected about $170 billion for calendar 2026, while Meta projected $135 billion to $145 billion, including finance-lease payments. These figures use company-specific definitions and should not be treated as directly comparable totals.
The source also compares capital spending with cash generated from operations in the quarter. Amazon reported $42.4 billion, Alphabet $42.3 billion, Meta $32.8 billion and Microsoft $58.0 billion in operating cash flow. The source calculates that only Microsoft had a positive remainder after its reported cash property and equipment spending. That comparison is a snapshot of one quarter, not a full measure of each company’s capacity to finance investment over time.
“Through a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.”
— Meta, in a statement on October 26, 2026
Who Carries the Long-Term Risk
The reported spending and financing figures establish the scale of investment, but they do not show how much revenue the new AI capacity will generate or when it will do so. It remains unclear whether demand will grow quickly enough to support the companies’ expected spending and the long-term payments attached to leases, power agreements and debt.
The approximately $1.02 trillion in unstarted lease commitments is a combined figure calculated from filings, and the source does not provide a company-by-company breakdown here. It also does not state how much of the total relates specifically to AI facilities, or how much can be reduced, renegotiated or assigned to another party. The potential effect on electricity rates is also not quantified.
The source frames the “bubble” debate as a question of who ultimately holds the risk. That is an interpretation, not a confirmed outcome: losses could fall differently across technology companies, lenders, investors, building owners and utilities, depending on contract details and future demand. No evidence in the supplied material establishes that a financing failure or consumer price increase has occurred.
Watch Spending and Commitments
The next signals will come from company earnings reports and filings, which can show whether projected 2026 spending is on track and how much is funded from cash versus debt, leases or partnerships. Investors and readers can also watch whether companies revise their 2027 capital plans as they report results and describe demand for AI services.
Further detail on lease schedules, power agreements and financing partners would help clarify how much cost remains with the technology companies and how much is carried elsewhere. Until those disclosures are available, the spending totals are clearer than the eventual returns or the final distribution of risk.
Key Questions
How much did the four companies spend in the third quarter?
About $202 billion on property and equipment from July through September 2026, according to the source’s sum of company cash flow statements. The comparable figure for the same quarter a year earlier was about $97 billion.
Who pays for the new data centers?
The companies use a mix of operating cash, bonds and loans, leases, partnerships and equity. The precise funding mix varies by company and project, and some arrangements place obligations with investors, lenders or building owners as well as the technology companies.
What does the $1.02 trillion lease figure mean?
It is the source’s combined estimate of future payments on leases that had been signed but had not started as of September 30, 2026, mostly for data centers. It is not an amount due all at once, and the supplied material does not give a company-by-company or AI-only breakdown.
Could data center investment raise household electricity bills?
That is possible in some circumstances, but the supplied material does not establish that household bills have risen because of these projects or quantify a likely effect. The outcome depends on power contracts, utility investment and local rate rules.
Is the AI data center boom a bubble?
The figures do not settle that question. They show a sharp rise in investment and substantial long-term commitments, but future AI revenue and demand remain uncertain. The source identifies the distribution of risk—not a confirmed collapse—as a central issue.
Source: ThorstenMeyerAI.com
