Big Tech Has Already Spent $1.1T on AI Infrastructure Since 2023; Plans $745B More in 2026 as Analyst Projects $1T+ for 2027
Summary
- • Google, Amazon, Microsoft, and Meta spent a combined $1.1T in capex from the start of the AI boom in 2023 through June 2026
- • The four hyperscalers plan to spend $745B on AI infrastructure in full-year 2026 alone — already unprecedented single-year scale
- • Evercore ISI projects the same four companies will exceed $1 trillion in combined capex in 2027 as AI investment accelerates
- • Analyst sees Meta and Amazon better positioned than Google amid capex headwinds; stocks await peak-intensity inflection for rally
Details
$1.1T spent 2023–June 2026
Google, Amazon, Microsoft, and Meta combined capex from the start of the AI boom through Q2 2026, per Financial Times reporting
$745B planned for full-year 2026
Combined capex plan for the four hyperscalers in 2026 alone — representing an acceleration from 2025 and making it one of the largest single-year corporate investment events in history
$1T+ 2027 AI Capex Projected
Evercore ISI projects Google and Amazon each at $300B+, Meta and Microsoft each at $200B in 2027 — totaling $1T+ before SpaceX and Oracle
Stock rally catalyst thesis
Analyst Mark Mahaney argues hyperscaler stocks will see a major run when peak capex intensity becomes visible and market focus shifts to return on AI investment
Meta and Amazon favored over Google
Mahaney sees Meta and Amazon as having a more positive setup than Google given current capex headwinds and relative revenue trajectory
Persistent consensus underestimate
Street consensus estimates have consistently underestimated hyperscaler capex, creating ongoing market uncertainty around the AI investment trajectory
Actual historical capex data (FT) and analyst projections (Evercore ISI) for hyperscaler AI infrastructure spending 2023–2027
What This Means
The confirmation that just four tech companies have already spent $1.1 trillion on AI infrastructure since 2023 — and plan to add another $745 billion in 2026 alone — transforms what were once forward projections into documented historical fact. This is no longer a question of whether the investment will happen; it has happened and is accelerating. With Evercore ISI projecting $1 trillion-plus for 2027, the investment cycle is still building rather than plateauing. For markets, the critical question shifts from 'will they really spend this much?' to 'when will the returns justify this capital deployment?' — a question that will determine hyperscaler valuations for the next several years. For the broader economy, $1.1 trillion in AI infrastructure spending in roughly three years represents one of the largest coordinated capital allocation events in modern corporate history, with sweeping implications across energy, semiconductors, real estate, and labor markets.