AI capital expenditure is now large enough to move semiconductor revenue, power markets, construction capacity and bond issuance. The useful way to track it is company by company, keeping disclosed spending plans separate from industry estimates and distinguishing annual capex from multi-year project announcements.
What the evidence establishes
Alphabet raised its 2026 capital-expenditure forecast in July to $195 billion-$205 billion, with infrastructure demand still exceeding installed capacity. Reuters reported that Meta expected to spend as much as $145 billion on AI infrastructure in 2026 while targeting seven gigawatts of compute infrastructure during the year. Google separately announced at least €13 billion of AI infrastructure investment in Finland over two years. Microsoft has not provided a directly comparable annual figure in the cited report, but a reported long-range plan targets approximately 38GW of data-centre capacity by 2032.
The commercial reading
These numbers should not be added mechanically. Alphabet's annual capex covers more than a single AI programme; Meta's figure is described as AI infrastructure; Google's Finland commitment spans two years and overlaps the broader Alphabet capital programme; Microsoft's capacity target is measured in gigawatts rather than dollars. The tracker is useful precisely because it preserves those definitions rather than manufacturing a false total.
What to watch next
Update this page after quarterly earnings and major project announcements. The most important changes are upward or downward capex revisions, new long-term power contracts, data-centre commissioning, depreciation growth and evidence that cloud or AI revenue is keeping pace with the installed asset base.
How to use this analysis
Source and verification note
The reporting base for this article is Reuters: Alphabet raises 2026 capex forecast and Reuters: Meta AI chips and 2026 infrastructure spending and Reuters: Google €13bn Finland AI infrastructure investment and Reuters: Microsoft data-centre capacity plan. The link is provided to the source page or release so readers can check the reporting period, definitions and later revisions. Figures are not extended beyond the source's geographic or institutional scope, and forecasts remain labelled as expectations until an official release records the outcome.
| Company | 2026 / current disclosure | What it measures |
|---|---|---|
| Alphabet | $195bn-$205bn 2026 capex forecast | Company-wide capital expenditure, heavily driven by AI/cloud infrastructure |
| Meta | Up to $145bn AI infrastructure spending in 2026 | AI infrastructure; Reuters-cited internal plan |
| Google Finland | At least €13bn over two years | Three new Finnish data centres and associated infrastructure |
| Microsoft | ~38GW data-centre capacity target by 2032 | Reported long-range capacity target, not annual capex |