🔍 Read the full analysis: Who Is Financing The Race To Build AI Data Centers? on ThorstenMeyerAI.com
Get the little things that make your day 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 their combined spending a year earlier, according to figures compiled from company filings. The companies are paying through operating cash, debt, leases, property partnerships and share sales, while the total and allocation of future commitments—and the returns from AI—remain 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 spent in the same quarter a year earlier, according to a calculation from their cash flow statements. The companies are funding the AI infrastructure surge with a mix of operating cash, borrowing, leases, property partnerships and stock sales—spreading some costs and risks beyond their own balance sheets.
The financing shift was visible in several transactions this week. Meta sold $25 billion of common stock and $12.5 billion of preferred shares, according to the source report. Microsoft issued $25 billion in bonds, including debt with a term extending to 2066. Amazon disclosed Friday that it had arranged a loan of up to $80 billion, secured by servers it will purchase with the proceeds. The arrangements differ, but each gives the companies access to capital for a build-out that is growing faster than their recent equipment spending alone.
Company filings cited in the report show the scale of the quarterly outlays: 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’s figure rises to $45.2 billion when new finance leases are included. The companies’ stated 2026 spending expectations also remain high: Amazon expects about $220 billion in cash capital spending, Alphabet $195 billion to $205 billion, Microsoft about $170 billion, and Meta $135 billion to $145 billion, with each figure reflecting the company’s own accounting or stated scope.
Cash generation does not cover the full investment for most of the group in the quarter. Amazon, Alphabet and Meta each spent more on property and equipment than the cash their operations generated, based on the source’s comparison; Microsoft generated about $58 billion in operating cash and spent $39.4 billion in cash capital expenditure. The report also calculates that the four companies had about $1.02 trillion in future lease payments for leases that had not yet started as of September 30. That total comes from adding figures in their filings and is not the same as debt already due.
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 Infrastructure Costs
The financing mix matters because it determines who is exposed if AI revenue grows more slowly than infrastructure costs. Borrowing leaves a company with scheduled interest and repayment obligations. A lease commits it to future payments, while a partnership or separate property owner can put buildings and financing on another entity’s books. Selling shares raises money without adding the same kind of debt repayment obligation, but dilutes existing shareholders’ ownership.
The source report identifies a timing mismatch: servers are generally written off over five to six years, while some power contracts run 15 to 20 years, data center leases can last up to 20 years, and Microsoft’s new bond issuance includes debt reaching 40 years. Those different timelines do not by themselves establish a loss, but they mean companies may be committed to costs after particular equipment has aged or been replaced. If demand and revenue fall short of plans, the financial pressure could extend to lenders, investors, building owners and utilities.
The immediate stakes are not limited to shareholders. Large data center projects also require electricity and long-term power arrangements. The source argues that some risk can reach electricity customers, but it does not quantify any pass-through to household bills or establish that consumers will bear a specific cost. The distinction matters: potential exposure is not evidence that a cost has already been transferred.
Products Worth Considering
Oura Ring 5 Smart Ring, Silver, Size 7
Levoit Vital 100S-P Air Purifier with Washable Pre-Filter
Garmin Forerunner 165 GPS Running Smartwatch, 43 mm, Black
As an affiliate, we earn on qualifying purchases.
From Cash Funding to Mixed Finance
The source describes a change from the prior pattern, in which major technology companies could fund data center construction from profits generated by advertising, software and online commerce. In 2026, the scale of planned AI infrastructure spending has prompted the companies to use multiple forms of financing at once. The source’s account of debt raised by Amazon, Alphabet and Meta puts their combined long-term borrowing at about $225 billion in the first nine months of the year; Microsoft’s $25 billion bond sale came later in October.
Meta set out its intended approach in a statement on October 26, saying its AI investments would be funded through “a balanced mix” of operating cash flow, investment-grade debt, strategic partnerships and equity capital. That is a company description of its plan, not a guarantee about the eventual amount or source of funding. The source also points to newer GPU cloud providers, including CoreWeave, whose senior notes it says carry interest rates of 8.5% to 9.75%. Those rates illustrate a different financing profile from the large technology companies, but the source does not provide a full comparison of those companies’ total obligations or ability to repay.
Returns and Final Risk Holders
The spending and financing figures are tied to company filings and disclosures, but the future return on the investment is not yet established. The source does not provide a common measure of AI revenue or profit across the four companies, nor does it show how quickly each project is expected to earn back its cost. Company spending forecasts can also change as construction schedules, equipment orders and demand shift.
The roughly $1.02 trillion lease figure is the source author’s sum of four filings, covering future payments for leases that had not started by September 30. It should not be read as $1.02 trillion of current borrowing, immediate payments or an estimate of losses. The source does not break down every commitment by project or explain which counterparties ultimately bear each risk. It also does not quantify how financing costs might affect data center prices, electricity rates or pension funds.
Company Spending Plans Through 2026
The next markers are the companies’ actual spending and financing disclosures against their 2026 plans. Amazon, Alphabet, Microsoft and Meta have each stated high capital spending expectations for the year, but the figures are forecasts, not completed spending. Their subsequent filings should show how much was paid in cash, added through leases or funded by new borrowing and equity.
Investors and readers can also watch whether AI-related revenue and operating cash flow rise alongside infrastructure costs, and whether companies revise their spending plans for 2027. For now, the confirmed development is a financing mix widening as the build-out accelerates; how profitable the facilities will be, and which parties ultimately absorb any shortfall, remains open.
Key Questions
Who is financing the AI data center build-out?
The companies themselves and outside capital providers are financing it. The mix includes operating cash, bonds and loans, leases, partnerships with property owners, and equity sales. Investors and lenders provide capital through some of those channels; the source does not identify every counterparty.
How much did the four companies spend in the third quarter of 2026?
Amazon, Alphabet, Microsoft and Meta spent about $202 billion combined on property and equipment from July through September 2026, according to a sum of their cash flow statements in the source report. The measure covers property and equipment, not only data centers or AI-specific purchases.
Does the $1.02 trillion lease figure mean the companies owe that amount now?
No. The reported $1.02 trillion is a sum of future payments for leases that had not started by September 30. It is a measure of future commitments, not a claim that the full amount is currently due or is equivalent to debt already borrowed.
Are the companies’ AI investments guaranteed to pay off?
No return is guaranteed. The source reports spending, financing and company forecasts but does not establish how much revenue or profit the infrastructure will generate. The pace of AI demand and the facilities’ eventual utilization remain uncertain.
Source: ThorstenMeyerAI.com
