🔍 Read the full analysis: Who Pays For The AI Data Center Boom? A Guide To The Money on ThorstenMeyerAI.com
Get movie nights 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 the third quarter of 2026, more than twice the year-earlier total. They are financing the build-out with operating cash, borrowing, leases, partnerships and equity, while long-term commitments shift some risk to investors, landlords and power customers.
Amazon, Microsoft, Alphabet and Meta spent about $202 billion on property and equipment from July through September 2026, more than twice their roughly $97 billion total a year earlier, according to a calculation from their cash-flow statements in a report published October 30. The companies are financing the spending with cash, debt, leases, partnerships and stock sales, creating long-term obligations alongside their investments in equipment and facilities.
The companies’ reported quarterly spending was $77.3 billion for Amazon, $51.4 billion for Alphabet, $39.4 billion for Microsoft and $33.9 billion for Meta. Microsoft’s figure rises to $45.2 billion when new finance leases are included, according to its filing. The totals cover property and equipment, not data centers alone; Amazon said its increase primarily reflected AI investment, while its spending also includes warehouses and other assets.
The cash-flow comparison helps explain the financing shift. In the quarter, the four companies generated operating cash of $42.4 billion at Amazon, $42.3 billion at Alphabet, $58 billion at Microsoft and $32.8 billion at Meta. Those figures are not the same as cash available after all expenses. The source report calculates that only Microsoft remained clearly cash-flow positive after property and equipment spending, while Amazon, Alphabet and Meta exceeded their quarterly operating cash on that measure.
They are turning to capital markets and outside financing. The report says Amazon sold $25 billion of stock and $12.5 billion of preferred shares during the week; Microsoft issued $25 billion in bonds, including debt extending to 2066; and Amazon disclosed a loan facility of up to $80 billion secured by servers it plans to buy. It also calculates that the four companies had about $1.02 trillion in future lease payments for leases not yet started, based on their September 30 filings. That is a report-derived total, not a sum stated by the companies as one combined figure.
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 Long-Term Cost
The financing arrangements allocate financial obligations among the companies and other parties. Bondholders lend directly to the companies, while leases and partnerships can place buildings or other assets on separate balance sheets. Companies remain responsible for payments under their contracts, and investors or property owners may face losses if projects do not generate sufficient returns.
The source report says servers are typically written off over five to six years, while data-center leases may run as long as 20 years and power contracts 15 to 20 years. Some bonds extend farther. If computing demand or AI income grows more slowly than expected, companies may still owe money for facilities and electricity after the equipment has lost value or become outdated.
The report identifies newer GPU cloud providers, including CoreWeave, and companies that own data-center buildings as potentially more financially exposed. It cites senior notes at CoreWeave carrying rates of 8.5% to 9.75%. These rates illustrate differences in financing costs across the sector but do not establish that losses will occur. Investors, pension funds and electricity customers may also be affected through their investments or the infrastructure and power markets they rely on.
Popular right now
Bose QuietComfort Headphones (2nd Gen), Wireless Noise-Cancelling Over-Ear Headphones, Black
Dyson Car+Boat Cordless Handheld Vacuum
KitchenAid Artisan 5-Qt Tilt-Head Stand Mixer – Almond Cream (KSM150PS)
As an affiliate, we earn on qualifying purchases.
The Spending Behind the Financing
For years, large technology companies could fund facilities from profits generated by advertising, software and online commerce. The current spending pace is much larger: the report’s calculation puts combined quarterly capital spending at about $202 billion in Q3 2026, compared with approximately $97 billion in the same period a year earlier. Each company’s total includes property and equipment beyond AI infrastructure, so the combined figure should not be read as an exact measure of AI-only spending.
Company guidance points to continued high investment. Amazon projected about $220 billion in cash capital spending for 2026 and said the increase primarily reflects AI. Alphabet kept its 2026 capital-spending outlook 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 company forecasts use different definitions and should not be treated as directly comparable estimates of the same spending category.
The report describes five funding routes: operating cash, borrowing, leases, partnerships or separate asset-owning companies, and equity. Power contracts add another long-term commitment. Meta stated on October 26 that it intends to fund AI investment through “a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.” That is the company’s stated approach; it does not specify how much will come from each source.
““Through a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.””
— Meta, in a statement cited in the October 30 report
Revenue Against Contracted Costs
The filings and company statements show spending, financing and disclosed commitments; they do not establish how much revenue the new AI capacity will generate. Future demand, utilization and returns remain uncertain, as do the eventual cost and timing of projects still under development. The combined $1.02 trillion lease figure is the report author’s sum of four filings and refers to future payments on leases that had not started by September 30.
The figures also use varying accounting categories and periods. Microsoft’s stated capital spending includes finance leases in one presentation, while the other company figures cited are cash property-and-equipment spending. Forecasts are company guidance, not completed spending. The report’s calculations and claims should be distinguished from figures explicitly presented as combined totals by the companies themselves.
It is not yet clear how much of the financing burden will ultimately sit with the technology companies, lenders, landlords, infrastructure partners or utility customers. Nor do the cited commitments alone establish that the sector is in a bubble. The relevant test is whether revenues and asset use can support the long-term payments, and that cannot be determined from quarterly spending figures alone.
The Next Spending and Filing Tests
The next evidence will come from subsequent quarterly filings and company updates on capital spending, cash flow, debt and lease obligations. Investors will be watching whether the 2026 spending forecasts hold, how much additional investment companies project for 2027, and whether AI-related revenue and demand keep pace.
Further disclosures may clarify financing terms, including the final amount and conditions of Amazon’s reported loan facility, as well as the roles of partners and property-owning entities. The cited material does not provide a timetable for resolving those questions. Until companies report more, the clearest measure is the gap between spending and internally generated cash, alongside the duration and cost of the obligations they take on.
Key Questions
How much did the four companies spend in the third quarter of 2026?
About $202 billion on property and equipment from July through September, according to the report’s sum of their cash-flow statements. The total is not limited to AI data centers.
Who is paying for the data center expansion?
The companies are using a mix of operating cash, bonds and loans, leases, partnerships and equity. The specific mix differs by company and is not fully disclosed in the cited material.
What does the $1.02 trillion lease figure mean?
It is the report author’s combined estimate of future payments on leases, mostly for data centers, that had not started by September 30, 2026. It is not a combined figure directly reported by the four companies.
Does the spending prove there is an AI bubble?
No. Spending and long-term commitments show the scale of investment and financial exposure, but they do not prove that projects will fail or that current valuations are unsustainable. Future AI revenue and facility use remain uncertain.
Source: ThorstenMeyerAI.com
