AI Statistics CenterAI Statistics Center

Market & Investment

How much are hyperscalers spending on AI in 2026?

By AI Statistics Center Editorial TeamLast updated: Reviewed against primary sources

Short Answer

Wall Street consensus for 2026 hyperscaler AI capital spending is $527 billion, with total global data-centre construction projected at $2.9 trillion through 2028 (Goldman Sachs, Morgan Stanley, 2026).

Key Facts

  • $527B, consensus estimate for 2026 hyperscaler AI capex (Goldman Sachs Research, 2026).
  • $2.9T, in global data center construction projected through 2028 (Morgan Stanley Research, 2026).
  • $500B, in AI-related spending projected for 2026 (UBP Investment Outlook 2026, 2026).
$527B

consensus estimate for 2026 hyperscaler AI capex

Wall Street consensus for 2026 hyperscaler capital spending is now $527 billion, up from $465 billion at the start of Q3 2025 earnings season, continuing a trend of upward revisions.

Who are the biggest AI spenders?

Capital expenditure from Microsoft, Alphabet, Amazon, and Meta is expected to rise by more than 34% again in 2026 to about $500 billion combined. Microsoft and Amazon lead on absolute dollars; Meta leads on percentage growth; Alphabet leads on return-on-ad-spend from AI.

How much of US GDP is AI investment?

Morgan Stanley estimates AI-related investment now accounts for around 25% of US GDP growth. The multiplier effect, construction, power, cooling, logistics, chips, is why the AI buildout is being compared to railroads and electrification rather than to prior software cycles.

Is the data-centre buildout near its peak?

No. More than 80% of the projected $2.9 trillion in data-centre construction through 2028 is still ahead, a multi-year industrial buildout, not a speculative tech cycle. Power availability and grid interconnect, not chip supply, is now the binding constraint.

Supporting Data

Recommended Citation

AI Statistics Center, citing Goldman Sachs Research (2026). https://aistatisticscenter.com/answers/how-much-are-hyperscalers-spending-on-ai

Free to quote with attribution. Linking back to the answer page helps readers verify the data and keeps this resource free.

Related Questions

Last reviewed and updated: by the AI Statistics Center Editorial Team. All statistics are sourced from primary research publications and linked directly to their origin.