🔍 Read the full analysis: Tracing The Costs Of The AI Data Center Boom on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
Amazon, Microsoft, Alphabet and Meta spent a combined $202 billion on property and equipment in the three months through September 2026, more than twice the year-earlier total, according to an analysis of their filings. The companies are financing the AI build-out with cash, debt, leases, partnerships and equity, while the scale and duration of future commitments raise questions about who bears the cost if AI revenue falls short.
Amazon, Microsoft, Alphabet and Meta spent about $202 billion on property and equipment in the three months through September, more than double the comparable amount a year earlier, according to an October 30 analysis of their company filings. As spending on AI data centers, servers and chips expands, the firms are combining operating cash with borrowing, leases, partnerships and share sales to pay for it.
The analysis, which adds figures from the companies’ cash flow statements, puts their combined third-quarter spending at about $202 billion, compared with about $97 billion a year earlier, amid a growing focus on keeping AI work clear and traceable. Amazon reported $77.3 billion in property and equipment purchases, Alphabet $51.4 billion, Microsoft $39.4 billion in cash spending and Meta $33.9 billion. Microsoft’s figure rises to $45.2 billion when new finance leases are included.
The article says three of the four firms spent more on property and equipment in the quarter than their businesses generated in operating cash. Microsoft generated $58 billion in operating cash, leaving about $18.6 billion after its cash property and equipment spending, based on the analysis’s arithmetic. Amazon generated $42.4 billion, Alphabet $42.3 billion and Meta $32.8 billion. These comparisons do not account for every financing source or use of cash.
Borrowing and other funding are taking a larger role. The source reports that Meta sold $25 billion of common stock and $12.5 billion of preferred shares during the week, Microsoft issued $25 billion in bonds, and Amazon disclosed a loan facility of up to $80 billion secured by servers it will buy. It also estimates that the four companies had about $1.02 trillion in future lease payments for leases not yet commenced as of September 30, based on their filings, commitments that make data center equipment planning increasingly consequential.
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
Who Carries the AI Build-Out Costs
The spending matters beyond the companies’ quarterly results because data centers require long-lived buildings, equipment and electricity commitments. The source analysis says servers are typically depreciated over five to six years, while data center leases may last up to 20 years and power contracts 15 to 20 years. Some bonds extend much longer: Microsoft’s reported bond sale reaches 2066.
That difference in timelines creates a potential financing mismatch. If AI-related income grows more slowly than expected, companies and their financing partners still face lease, debt and power obligations. The source argues that exposure may extend to investors, pension funds, data-center property owners and electricity customers, although it does not quantify how much risk each group holds.
The figures also complicate the debate over whether AI infrastructure is a bubble. Spending alone does not establish that investments will fail or that demand will not justify them. The sharper question is how the projects are financed, how quickly they generate revenue, and which parties absorb losses if returns disappoint.
Popular right now
Dyson Car+Boat Cordless Handheld Vacuum
Ninja Flip Toaster Oven and Air Fryer, 8-in-1, Stainless Steel, SP151
TCL QM64L 85-Inch QD-Mini LED 4K Smart Fire TV
As an affiliate, we earn on qualifying purchases.
From Cash-Funded Expansion to Mixed Finance
The source describes a shift from a period when large technology companies could fund data-center expansion from profits generated by advertising, software and online commerce. In 2026, the scale of planned construction and equipment purchases has led them to use several sources of capital at once. The five methods it identifies are operating cash, debt, leases, partnerships or separate property-owning entities, and equity issuance; long-term power contracts add further obligations.
Company forecasts cited in the article show the planned scale. Amazon expected about $220 billion in cash capital spending for 2026 and said the increase primarily reflected AI investment. Alphabet kept its 2026 capital spending forecast at $195 billion to $205 billion and anticipated a significant increase in 2027. Microsoft expected about $170 billion for calendar 2026, while Meta forecast $135 billion to $145 billion, including finance-lease payments. These are company projections, not completed spending totals.
The source also points to newer GPU cloud providers such as CoreWeave, whose senior notes it says carry interest rates of 8.5% to 9.75%. That example illustrates that financing terms can vary across the sector; it does not establish the overall risk profile of every AI infrastructure company.
“through a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.”
— Meta, in a statement dated October 26, as quoted by ThorstenMeyerAI.com
Revenue, Risk and Lease Exposure
The spending and financing figures are drawn from company filings, but the combined totals are the author’s calculations rather than a reported figure from any one company. The article does not provide a breakdown of the estimated $1.02 trillion in future lease payments by company, facility or expected start date, nor does it establish how much of that total is specifically tied to AI rather than other operations.
It is also unclear how much revenue the new infrastructure will generate, when facilities will reach full use, or how quickly demand for AI services may change. Forecasts for 2026 and 2027 are company guidance and may change. The source presents the possibility that some risk may sit with lenders, investors, property owners and power customers, but does not calculate potential losses or show that those parties will necessarily bear them.
Spending Plans and Filings Ahead
The next evidence will come from the companies’ upcoming filings and results, which can show whether spending continues at the projected pace and how it compares with operating cash generation. Investors and readers can also track whether the firms revise capital spending guidance, disclose additional borrowing or partnerships, or provide more detail about leases and financing arrangements.
For now, the figures establish the scale of the build-out and the growing mix of funding sources, not whether the investments will earn adequate returns. The central issue remains whether AI-related income can support obligations that may extend well beyond the useful life of some of the equipment being purchased.
Key Questions
How much did the four companies spend in the third quarter of 2026?
ThorstenMeyerAI.com’s calculation from company cash flow statements puts Amazon, Microsoft, Alphabet and Meta’s combined property and equipment spending at about $202 billion from July through September, up from about $97 billion in the same period a year earlier.
Does the $1.02 trillion figure mean the companies have borrowed that amount?
No. The source describes it as an estimated total of future payments on leases that had not started as of September 30, 2026. It is not the same as current borrowing or debt on the companies’ balance sheets.
Why are the companies using more than one way to finance data centers?
The article says the scale of construction and equipment spending is prompting firms to combine operating cash with bonds and loans, leases, partnerships and equity. The source does not claim that all four companies use each method to the same extent.
Does the spending prove that AI infrastructure is a bubble?
No. Large spending and long-term commitments show the scale of the investment, but do not by themselves prove that it will fail. The unresolved issue is whether future AI revenue will support the costs and who would bear a shortfall.
Source: ThorstenMeyerAI.com
