Do these companies report AI capex separately?

No. All four report total capital expenditure and attribute its growth to AI and cloud infrastructure in commentary. Any 'AI capex' figure you see elsewhere is an estimate.

How fast is this spending growing?

The combined quarterly flow (big-4 capex plus NVIDIA data-center revenue) went from $39.4B in Q1 2022 to $255.0B in Q2 2026 — about 55% a year compounded, with a dip in early 2023.

What was guided for next year?

Meta guided 2026 capex of $130–145B in its Q2 2026 release. Alphabet and Amazon capex guidance ($195–205B and ~$220B) appeared in earnings-call coverage rather than the releases, so those two are weaker-sourced.

BRAIAIN Capital Last updated: 2026-10-20

How much are hyperscalers spending on AI?

Name: Big-4 hyperscaler quarterly capex + NVIDIA data-center revenue, 2022–2026
Creator: BRAIAIN
Published: 2026-10-20
License: https://creativecommons.org/licenses/by/4.0/

In the four quarters ending June 30, 2026, Microsoft, Alphabet, Amazon and Meta reported $513.8 billion of combined capital expenditure — Amazon $173.0B, Alphabet $132.4B, Microsoft $115.9B, Meta $92.4B. The companies describe the growth as AI- and cloud-driven in commentary, but none reports an AI line item, so these are total-capex figures as filed.

BRAIAIN Buildout Index: big-4 capex + NVIDIA data-center revenue per calendar quarter, Q1 2022 = 100, scale 0–648 (overlapping lines summed; see method). Latest: 647.7. Chart drawn from the table below.

The data

quarter	msft capex musd	googl capex musd	amzn capex musd	meta capex musd	nvda dc revenue musd	combined flow musd	index q1 2022 eq 100
2022Q1	5340	9786	14951	5548	3750	39375	100.0
2022Q2	6871	6828	15724	7747	3806	40976	104.1
2022Q3	6283	7276	16378	9518	3833	43288	109.9
2022Q4	6274	7595	16592	9223	3616	43300	110.0
2023Q1	6607	6289	14207	7087	4284	38474	97.7
2023Q2	8943	6888	11455	6354	10323	43963	111.7
2023Q3	9917	8055	12479	6763	14514	51728	131.4
2023Q4	9735	11019	14588	7899	18404	61645	156.6
2024Q1	10952	12012	14925	6715	22563	67167	170.6
2024Q2	13873	13186	17620	8472	26272	79423	201.7
2024Q3	14923	13061	22620	9202	30771	90577	230.0
2024Q4	15804	14276	27834	14836	35580	108330	275.1
2025Q1	16745	17197	25019	13692	39112	111765	283.8
2025Q2	17079	22446	32183	17012	41096	129816	329.7
2025Q3	19394	23953	35095	19370	51200	149012	378.4
2025Q4	29876	27851	39522	22140	62300	181689	461.4
2026Q1	30876	35674	44203	19840	75200	205793	522.6
2026Q2	35802	44924	54208	31080	89000	255014	647.7

Download: CSV · JSON — compiled data CC BY 4.0; underlying sources keep their own terms.

Method & caveats

Each company's own headline capex definition is used as filed: Microsoft and Alphabet report cash purchases of property and equipment; Amazon reports gross purchases (not netted for proceeds); Meta includes finance-lease principal, matching the figure its releases state. Quarters not reported directly are derived by subtracting year-to-date figures within or across the company's own filings; every derivation is recorded in the data file. The index column adds NVIDIA's data-center segment revenue, whose fiscal quarters end about a month after calendar quarters; chips the big 4 buy from NVIDIA appear in both components, so the combined flow is a sum of reported lines with the overlap not removed — it is labeled accordingly wherever it appears.

Sources

Questions

Do these companies report AI capex separately?: No. All four report total capital expenditure and attribute its growth to AI and cloud infrastructure in commentary. Any 'AI capex' figure you see elsewhere is an estimate.
How fast is this spending growing?: The combined quarterly flow (big-4 capex plus NVIDIA data-center revenue) went from $39.4B in Q1 2022 to $255.0B in Q2 2026 — about 55% a year compounded, with a dip in early 2023.
What was guided for next year?: Meta guided 2026 capex of $130–145B in its Q2 2026 release. Alphabet and Amazon capex guidance ($195–205B and ~$220B) appeared in earnings-call coverage rather than the releases, so those two are weaker-sourced.

Cite this

BRAIAIN. "How much are hyperscalers spending on AI?" www.braiain.com/hyperscaler-capex.html, updated 2026-10-20. Original sources as listed above.

Part of BRAIAIN — the AI buildout, measured. Every figure carries its source and date; the full rules are in the methodology. Corrections: see the contact note there. © 2026 BRAIAIN, a dnilgis project.