TL;DR
Get backup power and energy gear delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Amazon, Microsoft, Alphabet and Meta spent a combined $202 billion on property and equipment in July–September 2026, more than twice the amount a year earlier. Their financing now combines operating cash, borrowing, leases, partnerships and stock, while long-term commitments raise questions about who bears the cost if AI revenues lag.
Amazon, Microsoft, Alphabet and Meta spent about $202 billion on property and equipment from July through September 2026, more than twice their combined spending a year earlier, according to calculations based on company filings. The pace is pushing even highly profitable companies to use debt, leases, partnerships and stock sales alongside their own cash to finance data centers, servers and chips.
The spending total is a calculation from the companies’ cash flow statements, and covers property and equipment rather than AI alone. Amazon reported $77.3 billion in purchases, Alphabet $51.4 billion, Microsoft $39.4 billion and Meta $33.9 billion. Amazon also builds warehouses, though the company said its increase primarily reflected investment in artificial intelligence. Microsoft’s quarter was its fiscal first quarter of 2027.
The reported figures point to pressure on cash flow. Amazon’s spending exceeded cash from operations by about $34.9 billion in the quarter. Alphabet reported quarterly free cash flow of negative $9.1 billion, and Meta reported negative $2.1 billion. Microsoft generated $58 billion in operating cash against $39.4 billion in property and equipment purchases. These measures have different accounting definitions, and the spending figures include projects beyond AI.
Companies are drawing on several funding sources at once. The source material says Amazon, Alphabet and Meta raised about $225 billion in long-term debt during the first nine months of 2026. Microsoft sold $25 billion in bonds during the week of October 30, including debt with a 40-year term. Meta said on October 26 that it plans to fund AI investment through operating cash flow, investment-grade debt, strategic partnerships and equity capital. Amazon disclosed a loan facility of up to $80 billion secured by servers it plans to purchase.
Capital, commitments & accountability · Q3 2026
4 Ways The AI Data Center Boom Gets Funded—and Who Pays
Amazon, Microsoft, Alphabet and Meta spent a combined $202 billion on property and equipment in July–September 2026—more than twice a year earlier. Their funding mix reaches beyond cash, while commitments can last long after hardware ages.
01 / Spending at scale
Four balance sheets, one rapid build-out
The quarterly property and equipment figures below are based on company cash flow statements. They show the scale of investment, not a clean measure of AI spending.
Microsoft’s quarter was its fiscal first quarter of 2027.
Amazon
$77.3B Purchases; includes warehouses as well as data centers.Alphabet
$51.4B Purchases of property and equipment in the quarter.Microsoft
$39.4B Purchases against $58B in operating cash.Meta
$33.9B Purchases of property and equipment in the quarter.Combined quarterly purchases · approximate
The year-over-year comparison uses the same three-month period. Figures are the author’s sum of company filings and include projects beyond AI.
02 / The funding mix
Five sources help carry the build
Companies can combine funding sources rather than rely on one stream. Each shifts when cash is raised, and who is exposed if expected returns fall short.
Operating cash
Cash generated by advertising, software, cloud and online retail can fund investment directly.
Internal fundingBorrowing
Amazon, Alphabet and Meta raised about $225B in long-term debt in the first nine months of 2026. Microsoft sold $25B in bonds, including 40-year debt.
Lenders & bondholdersLeases
Facilities can be leased rather than owned. Unstarted lease payments are future commitments, not cash already spent.
Long-term paymentsPartnerships
Strategic partners and data center owners can share the financing and operating arrangements.
Shared exposureEquity
Stock sales or equity capital bring in funding while spreading ownership and potential returns.
Investor capital03 / Cash pressure & commitments
Spending is outrunning cash at some firms
Quarterly cash measures differ by company and accounting definition. They offer context for financing needs, but are not directly interchangeable.
Cash flow signals
- AmazonPurchases exceeded operating cash by about $34.9B.
- AlphabetReported free cash flow of negative $9.1B.
- MetaReported free cash flow of negative $2.1B.
- Microsoft$58B operating cash against $39.4B in purchases.
04 / Who may bear the cost?
Risk follows the contracts—and the revenue
If AI demand or revenue growth misses expectations, payments may continue after equipment loses value or becomes outdated. The source material describes potential pressure; it does not establish that losses will occur.
Statements in the filings
Plans are signals, not guarantees
Company outlooks describe intended spending and financing. They do not guarantee eventual investment or returns.
“through a balanced mix of operating cash flow, investment-grade debt, strategic partnerships, and equity capital.”Meta · Statement dated October 26, 2026
“primarily reflects investments in artificial intelligence.”Amazon · Third-quarter 2026 results
“significant”Alphabet · Expected spending increase in 2027
What to watch next
Match spending with evidence of returns
The build-out’s scale is visible. How much it earns—and who absorbs any shortfall—remains unsettled.
Forecasts and quarterly filings
Compare future spending updates with Amazon’s 2026 outlook, Alphabet’s 2026 guidance and expected 2027 increase, and financing disclosures from Microsoft and Meta.
Evidence that the assets are earning
Track revenue from AI products and cloud services, alongside facility utilization, power prices and the performance of specialized equipment.
Who pays, and when?
More detail on lease schedules, borrowing terms and data center partnerships would clarify how commitments are distributed over time.
Separate AI from total capital spending
The reported totals include non-AI projects. Amazon’s purchases, for example, also include warehouse construction; company filings do not quantify AI’s share across the four.
Quick answers
Three questions behind the numbers
How much did the four companies spend?
About $202B on property and equipment from July through September 2026, based on a sum of cash flow statements. The figure includes spending beyond AI.
What are the main funding sources?
Operating cash, borrowing, leases, partnerships and equity, including stock sales. Long-term power agreements also shape future costs.
What does the $1.02T lease estimate mean?
Estimated future payments on leases that had not started by September 30, mostly for data centers, summed from four company filings. It is not cash already spent.
Who Carries the Long-Term Costs
The financing mix matters because the assets and contracts supporting the build-out last for different lengths of time. Servers are generally written off over five to six years, while some power agreements run for 15 to 20 years, data center leases can extend up to 20 years, and some bonds mature after 40 years. If demand for AI services or revenue growth falls short of expectations, companies may still owe payments after the equipment has lost value or become outdated.
The immediate exposure varies. The largest technology companies have substantial businesses and can absorb investment losses, but risk also reaches lenders, bondholders, investors in companies that own data centers, and electricity customers. The source material identifies newer GPU cloud providers such as CoreWeave as more exposed to financing costs; it cites senior notes carrying rates of 8.5% to 9.75%. Those rates do not alone establish whether a company can meet its obligations, but they show that borrowing costs differ across the sector.
This makes the debate over an AI investment bubble partly a question of where financial risk sits. A slowdown could affect company returns, the value of specialized buildings and equipment, and the economics of long-term power commitments. The source material does not establish that these losses will occur; it describes potential pressure if revenue growth fails to support the spending.
Products Worth Considering
Blueair Blue Pure 211i Max Smart Air Purifier
Sonos Era 100 SL Compact, Microphone-Free Speaker – Black
Shark HydroVac MessMaster Corded Wet/Dry Vacuum, Mop & Self-Clean (WD161)
As an affiliate, we earn on qualifying purchases.
From Cash-Funded to Mixed Financing
For years, large technology companies could fund data center expansion from cash generated by advertising, software and online retail. The figures in the source material show a sharp change in scale in 2026: the four companies’ combined quarterly property and equipment purchases rose from about $97 billion a year earlier to about $202 billion in the July–September quarter. That comparison is for the same three-month period in successive years, based on the author’s sum of company filings.
Borrowing and stock sales are only part of the commitments. The companies also have signed leases, mostly for data centers, that had not started as of September 30. The source material estimates their combined future payments at about $1.02 trillion, based on four quarterly filings. Those payments are commitments over time, not cash already spent or debt recorded in the same way as bonds.
The disclosures offer different views of the build-out. Amazon said it expects about $220 billion in cash capital spending for all of 2026 and that the increase primarily reflects AI investment. Alphabet kept its 2026 capital spending outlook at $195 billion to $205 billion and said it expects spending to increase significantly in 2027. These are company statements about plans, not guarantees of eventual spending or returns.
“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 Must Match the Commitments
It is not yet clear how quickly AI services will generate enough revenue to justify the scale and duration of these investments. The spending totals combine AI projects with other property and equipment, and the source material does not quantify what share of each company’s total is specifically for AI. Amazon’s figure, for example, includes warehouse construction as well as data center investment.
The estimated $1.02 trillion in future lease payments is the author’s sum of company filings; the source material does not break down the amount by company, facility or payment year. Nor does it establish how much risk building owners, lenders or other partners ultimately retain. The commitments create potential exposure, but their eventual cost will depend on contract terms, utilization, power prices and business performance.
Watch Spending and Returns
The next indicators are the companies’ upcoming quarterly filings and updates to capital spending forecasts, alongside evidence of revenue from AI products and cloud services. Readers can compare those results with the plans already stated: Amazon’s 2026 spending outlook, Alphabet’s guidance for 2026 and its expectation of a significant 2027 increase, and Microsoft’s and Meta’s disclosures on financing.
Further detail on lease schedules, borrowing terms and data center partnerships would help show how payments are distributed over time. Until companies provide clearer breakdowns of AI-specific spending and returns, the scale of the investment is confirmed, but how much it will earn—and who will absorb any shortfall—remains unsettled.
Key Questions
How much did the four companies spend in the third quarter of 2026?
Amazon, Microsoft, Alphabet and Meta spent a combined about $202 billion on property and equipment from July through September 2026, based on a sum of their cash flow statements. The figure includes spending beyond AI.
What are the main ways companies are financing data centers?
The source describes five main sources: operating cash, borrowed money, leases, partnerships and equity, including stock sales. Long-term power agreements also shape the costs of operating the facilities.
What does the $1.02 trillion lease figure represent?
It is an estimate of future payments on leases that had not started by September 30, 2026, mostly for data centers, summed from four company filings. It is a future commitment, not money already spent.
Does the spending prove that AI data centers are a bubble?
No. The spending and financing commitments are reported figures, but they do not prove that investments will fail. The central uncertainty is whether future AI revenue will support costs that can extend across many years.
Source: Thorsten Meyer AI
