🔍 Read the full analysis: AI’s Data Center Boom: The Costs Behind The Construction on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Amazon, Microsoft, Alphabet and Meta spent about $202 billion on property and equipment in the third quarter of 2026, more than twice their combined spending a year earlier, according to company filings. Their financing now includes borrowing, leases, partnerships and equity, while future data-center lease payments add substantial obligations. How quickly AI revenues grow compared with long-term infrastructure costs remains uncertain.
Amazon, Microsoft, Alphabet and Meta spent about $202 billion on property and equipment in the third quarter of 2026, more than twice the roughly $97 billion they spent in the same period a year earlier, according to figures compiled from company filings. As AI data-center construction expands, the companies are using more than operating cash to fund it, including bonds, loans, leases and equity financing.
The quarterly total is based on cash-flow statements and covers property and equipment, including data centers, servers and chips; it is not a measure solely of AI spending. Amazon reported $77.3 billion in third-quarter property and equipment spending, Alphabet $51.4 billion, Microsoft $39.4 billion in cash spending and Meta $33.9 billion. Microsoft said its figure rises to $45.2 billion when new finance leases are included. The companies’ reported categories and accounting treatments are not identical.
The source’s calculation of the four companies’ cash from operations shows that spending exceeded this quarterly cash measure at Amazon, Alphabet and Meta. Microsoft generated $58 billion from operations and spent $39.4 billion in cash on property and equipment. The comparison is not a full measure of each company’s financial position: it does not include all cash flows, existing cash reserves, or the varied ways companies classify and finance assets.
Financing announcements underscore the shift. Meta said it sold $25 billion of common stock and $12.5 billion of preferred shares during the week covered by the report. Microsoft issued $25 billion in bonds, including debt with a maturity in 2066. Amazon disclosed that it had arranged a loan of up to $80 billion, secured by servers it will purchase with the proceeds. The source also calculates that Amazon, Alphabet and Meta raised about $225 billion in long-term debt during the first nine months of 2026.
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 Build-Out Costs
The funding mix matters because it determines who bears the cost if demand disappoints. A company that pays from current cash flow carries the immediate expense itself. Debt brings repayment obligations, while leases and partnerships can place assets or liabilities on different balance sheets. Share sales bring in capital but dilute existing shareholders’ ownership.
The report calculates that the four companies had about $1.02 trillion in future payments on leases, mostly for data centers, that had not started as of September 30. That total is based on the writer’s sum of the companies’ quarterly filings; it is not equivalent to debt due immediately. Still, long commitments can leave companies paying for facilities even if their use or revenue falls short of expectations. The consequences could reach investors and lenders, as well as communities and electricity customers if power demand and infrastructure costs rise.
The central financial question is not simply how much the companies spend, but whether revenue from AI services and products develops quickly enough to cover costs that last for years. The four businesses have large existing operations, but the source does not establish how much AI revenue will be generated by the new infrastructure or when.
Products Worth Considering
Apple MacBook Air 13-inch with M5 chip, 16GB memory, 1TB SSD, and 12MP Center Stage camera (Midnight)
Apple Mac mini Desktop Computer with M4 chip, 10‑core CPU and GPU, 24GB Unified Memory, 512GB SSD, Gigabit Ethernet
FLEXISPOT Mochi-Pro Zero-Wall Oversized Power Recliner, Beige
As an affiliate, we earn on qualifying purchases.
From Cash Funding to Mixed Finance
The source describes a change from a period when large technology companies could fund data-center construction from profits generated by advertising, software and online retail. In 2026, it says, the scale of investment has led companies to combine operating cash, borrowing, leases, partnerships and equity. Long-term power contracts add another layer of commitments beyond construction and equipment purchases.
Meta set out that approach in a statement on October 26, saying it intended to fund AI investments through “a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.” The companies’ spending plans also point to continued investment: Amazon forecast about $220 billion in cash capital spending for 2026; Alphabet maintained guidance of $195 billion to $205 billion; Microsoft expected about $170 billion for the calendar year; and Meta forecast $135 billion to $145 billion, including finance-lease payments. These are company projections, not final full-year results.
Investment is not limited to the largest platforms. The source identifies newer GPU cloud providers such as CoreWeave and separate companies that own data-center buildings as possible pressure points. It says CoreWeave senior notes carry rates of 8.5% to 9.75%. The source does not provide enough detail here to compare the terms or risk of those notes with the financing used by the larger companies.
“through a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.”
— Meta, in a statement dated October 26, 2026
Revenue Versus Long-Term Commitments
How profitable the new capacity will be remains unresolved. The figures provided show spending and financing, but do not establish how much revenue the data centers will generate, when that revenue will arrive, or how much of the spending is specifically for AI rather than other business needs. Amazon, for example, also builds warehouses, although it said the spending increase primarily reflected AI investment.
The lease total is a calculation from four company filings, and the article does not give a company-by-company breakdown of when each future payment falls due. It also does not quantify the value or duration of all power contracts, or detail how risks are divided among technology companies, landlords, lenders and partners. Without those terms and revenue outcomes, the scale of exposure cannot be fully assessed.
The article frames the “bubble” debate as a question of who holds the risk, but the available figures do not establish that an AI investment bubble exists. Nor do they show that any company will be unable to meet its obligations. Spending, forecasts and debt rates describe current commitments; they are not proof of future losses.
Upcoming Spending and Filings
The next major test will be whether the companies’ stated 2026 spending plans are reflected in their final results and whether they revise those plans as demand develops. Their future quarterly filings should provide updated figures for capital spending, cash generation, debt and lease obligations. Investors will also watch for disclosures on how AI-related services contribute to revenue.
Further details on Amazon’s server-backed loan, the terms of new borrowing and the structure of data-center partnerships could clarify how costs are allocated. For now, the confirmed picture is one of fast-rising infrastructure spending and a broader financing mix; the long-term returns and who ultimately bears any shortfall remain open questions.
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 a calculation based on Amazon, Alphabet, Microsoft and Meta filings. The total is not exclusively AI spending.
Why are technology companies borrowing to build data centers?
The source says the scale of investment has exceeded what some companies generated in operating cash during the quarter. Borrowing is one of several funding methods, alongside cash, leases, partnerships and equity; the filings do not mean every company is unable to fund construction from its broader resources.
What does the $1.02 trillion lease figure represent?
It is the source author’s combined calculation of future payments on leases, mostly for data centers, that had not started as of September 30, 2026. It is a long-term commitment figure, not a bill due immediately or a measure of debt alone.
Does the spending prove there is an AI bubble?
No. The figures document high investment and a wider mix of financing, but do not establish that a bubble exists. The key unresolved issue is whether future AI-related revenue will justify the costs and commitments.
Source: ThorstenMeyerAI.com
