🔍 Read the full analysis: Cash, Bonds, Leases: The Financing Behind AI Data Centers 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 third quarter of 2026, according to figures compiled from their filings. Their financing now combines operating cash, debt, leases, partnerships and equity, while a large volume of future lease payments and long-term power commitments leaves investors and other parties exposed to the build-out.
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, as the companies expand data centers and buy servers and chips for artificial intelligence. Their filings and recent financing announcements show that cash from operations is no longer the only funding source: debt, leases, equity and partnerships are also helping pay for the build-out.
The $202 billion estimate is a sum of the four companies’ cash-flow statements for the quarter ended Sept. 30, as reported by ThorstenMeyerAI.com. Amazon recorded $77.3 billion in property and equipment spending, Alphabet $51.4 billion, Microsoft $39.4 billion and Meta $33.9 billion. Microsoft’s figure rises to $45.2 billion when new finance leases are included. The figures are not a measure of AI spending alone; the companies also invest in other property and equipment.
The report says three of the four companies spent more on property and equipment in the quarter than their businesses generated in operating cash. Amazon reported $42.4 billion in operating cash flow, Alphabet $42.3 billion and Meta $32.8 billion, compared with their respective equipment spending. Microsoft generated $58 billion in operating cash flow against $39.4 billion in cash property-and-equipment spending. Amazon said its spending increase primarily reflected AI investment. The companies’ full-year 2026 capital-spending outlooks were about $220 billion for Amazon, $195 billion to $205 billion for Alphabet, about $170 billion for Microsoft and $135 billion to $145 billion for Meta; the measures are company guidance and may not use identical accounting definitions.
Financing announcements this week illustrate the shift. Meta sold $25 billion of common stock and $12.5 billion of preferred shares, while Microsoft issued $25 billion in bonds, including debt maturing in 2066. Amazon disclosed a loan of up to $80 billion secured by servers it plans to buy. The source also calculates that the four firms had about $1.02 trillion in signed leases not yet started as of Sept. 30, based on their quarterly filings. That is a future-payment commitment, not cash already borrowed or spent.
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 Risk
The financing mix matters because it determines who bears the cost if AI revenue does not grow as quickly as infrastructure spending. A company can fund a facility directly, borrow to build it, lease it, or work with a separate owner or strategic partner. Each arrangement places obligations and risks differently among the technology company, lenders, investors and building owners.
The timing of the commitments is another concern raised by the source. Servers may be written off over roughly five to six years, while data-center leases can run for as long as 20 years and power contracts for 15 to 20 years. Microsoft’s bond sale includes a 40-year maturity. If demand or revenue falls short, costs tied to buildings, electricity and financing may persist after equipment has lost value or become outdated. That mismatch does not establish that losses will occur, but it helps explain why the debate about an AI investment bubble is also about where the downside would land.
The exposure is not limited to the largest technology companies. The source points to newer GPU-cloud providers, building-owning companies and their investors as possible pressure points. It cites CoreWeave senior notes carrying interest rates of 8.5% to 9.75%. More broadly, data-center electricity demand and long-term power arrangements can also affect utilities and customers, though the source does not quantify those effects.
Products Worth Considering
Jackery Explorer 1000 v2 Portable Power Station
Anker SOLIX S2000 Portable Power Station, 2,010Wh, 1,500W
Bose SoundLink Flex (2nd Gen) Portable Bluetooth Speaker
As an affiliate, we earn on qualifying purchases.
From Cash-Funded Builds to Mixed Financing
For years, large technology companies could finance data centers from earnings generated by advertising, software and online commerce. The 2026 spending totals show a change in scale: the source’s calculation puts quarterly property-and-equipment spending at about $97 billion in the same quarter a year earlier, against about $202 billion in 2026. The comparison is based on the source’s sum of the companies’ cash-flow statements.
The companies are now using several funding channels at once: operating cash, bonds and loans, leases, partnerships, and equity issuance. Meta described its approach on Oct. 26 as a “balanced mix” of operating cash flow, investment-grade debt, strategic partnerships and equity capital. Leases and power agreements add future obligations that may not appear as current borrowing, while separate property companies can finance buildings that technology firms then occupy.
The report’s headline financing figures should be read with their stated scope. It says Amazon, Alphabet and Meta raised about $225 billion in long-term debt during the first nine months of 2026, and Microsoft then announced its $25 billion bond issue. The $1.02 trillion lease estimate covers commitments for leases that had not yet begun, not all lease payments or debt. Company filings provide the underlying figures, while the combined totals and comparisons are the source author’s calculations.
““Through a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.””
— Meta, in a statement on Oct. 26, as quoted by ThorstenMeyerAI.com
Revenue and Long-Term Costs Remain Unsettled
The disclosed figures show current spending and contractual commitments, but they do not establish how much of the investment will earn an adequate return. The source does not provide a comparable forecast of future AI revenue across the four companies, nor does it identify the expected utilization or returns for every facility and server purchase. Whether demand will justify the pace and duration of the spending remains uncertain.
The $1.02 trillion lease total is an author calculation from company filings, and its scope is future payments on leases that had not started by Sept. 30. It should not be treated as an immediate bill or as equivalent to debt. The article also identifies risks for GPU-cloud firms, property owners, investors and electricity customers, but does not quantify the losses or costs any group might ultimately face. The amount of spending that is specifically for AI, rather than other infrastructure, is not separately stated for every company.
Upcoming Guidance and Funding Disclosures
The next evidence will come through company filings, capital-spending guidance and financing announcements. Amazon’s stated 2026 cash capital-spending outlook is about $220 billion; Alphabet expects $195 billion to $205 billion, while Microsoft expects about $170 billion for calendar 2026 and Meta forecasts $135 billion to $145 billion. Those outlooks can change, and the companies use different definitions of capital spending, so later comparisons will need to preserve each measure’s scope.
Investors and readers can also watch whether operating cash flow keeps pace with equipment spending, how much new debt and equity the companies raise, and how lease and power commitments develop. The current disclosures do not resolve whether AI infrastructure returns will cover those obligations over time. That question will depend on future demand, revenue and the ability to use the assets productively.
Key Questions
How much did the four companies spend in the third quarter of 2026?
Amazon, Microsoft, Alphabet and Meta spent about $202 billion combined on property and equipment from July through September, according to a total calculated from their cash-flow statements by ThorstenMeyerAI.com. The amount includes investment beyond AI infrastructure.
Does the $1.02 trillion lease figure mean the companies owe that amount now?
No. It is the source author’s estimate of future payments on leases that had not started by Sept. 30, 2026. It is a long-term commitment, not cash already spent or a current borrowing balance.
Why are the companies raising debt and equity if they are profitable?
Their infrastructure spending has grown sharply. In the third quarter, three of the four companies spent more on property and equipment than they generated in operating cash, according to the figures cited in the report. Debt, share sales, leases and partnerships provide additional ways to fund investment.
What is the main risk identified in the report?
The report highlights a possible mismatch between the life of equipment and the length of financial commitments. Servers may be written off in five to six years, while some leases and power contracts last much longer. Whether that creates losses depends on future AI demand, revenue and asset use.
Source: ThorstenMeyerAI.com
