AI Capex From 4 Tech Giants to Exceed Japan's GDP by 2030, Goldman Says
Goldman Sachs updated estimates for hyperscaler AI capex from 2025-2030, saying Meta, Microsoft, Amazon and Alphabet are on track to spend a combined $5.3 trillion by end-2030. Goldman said this exceeds the GDP of Japan, the UK, India and France, and that total data center, power and computing spending could reach $7.6 trillion over five years. The bank expects private markets to fund more of the buildout; it also cited capex guidance of up to $725 billion this year for the four firms.
Reinforces a sustained high-capex AI investment profile; near-term focus is whether funding needs and ROI concerns keep pressuring the multiple.
Goldman estimates Meta’s hyperscaler AI capex will be part of ~$5.3T spending through 2030, with $725B capex guidance this year cited.
Modest volatility bias around capex/financing headlines rather than a clean directional catalyst.
Background
Goldman updated hyperscaler capex estimates for 2025-2030, focusing on four AI-heavy mega-cap platforms and the broader data-center/power/computing buildout.
Why it matters
The key tradable angle is whether investors treat the capex cycle as durable growth (cloud/AI monetization) or as a valuation overhang (ROI uncertainty and financing/dilution).
Market relevance
Provides a capex-cycle read-through for mega-cap AI investors, with Alphabet’s financing detail and Meta’s recent reaction offering the most direct sentiment hooks.
Market effects
Reinforces a multi-year data center/power/computing investment cycle, likely keeping AI infrastructure supply-chain demand bid while sustaining ROI skepticism.
Global capex comparison framing (vs Japan/UK/India/France GDP) underscores worldwide scale, but no single region-specific policy trigger is cited.
Large aggregate AI infrastructure spending (~$7.6T over five years) can influence global capex expectations across semis, networking, power, and construction.
Alternative perspectives
High capex may be partially offset by improving unit economics (higher AI attach rates, utilization) even if headline spending stays elevated.
The article is estimate-driven; actual capex timing, utilization, and monetization (ads, cloud margins, chatbot pricing) could diverge materially from the capex-only narrative.
Key entities
- analyst_firmGoldman Sachs
Updated estimates for hyperscaler AI capex from 2025 through 2030 and framed the spending scale vs national GDPs.
- companyMeta
Included as one of the four hyperscalers; article references recent stock reaction tied to AI chatbot monetization and capex intensity.
- companyAlphabet
Included as one of the four hyperscalers; article cites a planned $80B stock sale to fund AI infrastructure.
