Big Tech's Combined AI Infrastructure Spending Set to Reach $725 Billion in 2026

The four largest U.S. hyperscalers — Amazon, Alphabet, Meta, and Microsoft — collectively plan to spend roughly $725 billion on capital expenditures in 2026, up approximately 77% from around $410 billion in 2025, according to figures compiled from each company's own second-quarter earnings guidance, reported by Yahoo Finance.

What Happened

Each hyperscaler raised its own 2026 capital expenditure guidance during its most recent earnings report: Amazon to roughly $200 billion, Microsoft toward $190 billion, Alphabet to as much as $205 billion, and Meta to a range of $125–145 billion. All four figures represent significant increases over each company's original 2026 guidance issued earlier in the year.

What This Means

This level of spending is overwhelmingly directed at AI-specific infrastructure — GPU and custom AI chip capacity, new data centers, and the power generation needed to run them — rather than general-purpose cloud expansion. Industry analysis from Citigroup estimates global AI compute demand will require roughly 55 gigawatts of new power capacity by 2030, translating to an estimated $2.8 trillion in incremental global spending, with $1.4 trillion of that in the U.S. alone.

Why It Matters to Businesses

The sheer scale of this buildout — hundreds of billions of dollars committed by just four companies in a single year — is the clearest available signal of how central AI infrastructure has become to competitive strategy at the largest technology companies. For businesses building on any of these cloud platforms, it suggests continued rapid growth in available AI compute capacity over the medium term, even as near-term capacity remains tight.

Industry Impact

Hundreds of billions of dollars that would previously have gone toward share buybacks or other capital allocation are now directed at AI infrastructure, with power and cooling capacity — not chip supply alone — increasingly cited as the binding constraint on how fast this buildout can proceed.

What to Watch Next

Whether power grid capacity and permitting timelines become a more visible bottleneck than chip supply over the next several quarters, and whether any of the four hyperscalers moderate their spending pace if AI revenue growth doesn't keep pace with infrastructure investment.

---

Source: Yahoo Finance, compiled from company earnings reports