Meta, Google, and Microsoft Triple Down on AI Spending
Summary
Three US tech giants — Meta, Alphabet (Google) and Microsoft — reported record revenues while sharply increasing capital expenditure to scale AI infrastructure. Each company raised or confirmed multi‑billion‑dollar spending plans aimed at data centres, cloud capacity and AI tooling, even as analysts warn the sector could be heading toward a bubble.
Key Points
- Meta raised its annual capex guidance to $70–72bn and expects spending to be notably larger next year; Q3 revenue was $51.24bn (up 26% year‑on‑year).
- Alphabet now expects 2025 capex of $91–93bn (previously ~$75bn); Q3 revenue hit a record $102.3bn (up 33%).
- Microsoft reported $77bn in quarterly revenue (up 18%) and $34.9bn in capex this quarter — a 74% jump year‑on‑year, driven largely by AI infrastructure.
- Companies are investing heavily in data centres and making infrastructure more fungible and regularly modernised to follow rapid hardware advances.
- Massive outside investments and partnerships (eg. Nvidia/OpenAI, Microsoft/OpenAI) and announcements of multi‑year, multi‑gigawatt projects are stoking concerns of an AI spending bubble.
- OpenAI’s stated plans for 30GW of compute (and Nvidia’s conditional $100bn offer to OpenAI) highlight the scale and long‑term capital intensity of the race.
Content Summary
On the same earnings day, Meta, Alphabet and Microsoft all signalled they are accelerating investment in AI infrastructure. Meta is front‑loading capacity to be ready for potential breakthroughs and has been aggressively recruiting AI talent while reorganising teams to squeeze efficiency. Alphabet attributes its capex jump to data centres and AI initiatives, noting strong cloud growth and rapid user adoption of Gemini. Microsoft’s capex surge is tied to cloud and partnerships with frontier AI providers; it also recorded a one‑time net income hit related to its OpenAI investment and warned of ongoing volatility from that partnership.
Analysts point out that while building capacity in stages offers flexibility, the sheer scale of announced projects — from gigawatts of compute to multi‑billion dollar investments — raises the possibility of an overheated market. Companies argue the spend is necessary to meet demand and to remain competitive as models and hardware evolve quickly.
Context and Relevance
This is a useful snapshot of the current AI infrastructure arms race. The spending decisions here shape cloud pricing, enterprise AI availability, data‑centre construction booms, geopolitical supply chains for chips and energy demand. For investors, IT strategists and policymakers, the story highlights both long‑term bets that could enable new AI capabilities and short‑term financial risks if demand projections don’t pan out.
Why should I read this?
Because the big players are literally betting the farm on AI — if you care about where cloud costs, startups, or enterprise AI projects are heading, this explains why everything from hiring to real‑estate for data centres is suddenly weather‑sized. Also: bubble talk. Worth five minutes to know if you should be worried, excited or both.
Author style
Punchy. This piece gives a clear financial reality check: huge revenues paired with even huger infrastructure bets. If you follow tech markets or AI strategy, the details matter — this isn’t fluff.
Source
Source: https://www.wired.com/story/microsoft-google-meta-2025-earnings/
