Via datacenterknowledge.com
AI infrastructure demand is driving the biggest capital spending cycle in tech history, with major cloud providers committing hundreds of billions each
Six months ago, Wall Street expected the biggest cloud companies on Earth to spend around $460 billion on infrastructure in 2026. That number has now ballooned to $789 billion, a 71% revision that reflects just how aggressively Big Tech is betting on artificial intelligence.
The numbers behind the spending spree #
Goldman Sachs now estimates hyperscaler capex at roughly $800 billion for 2026, placing the bank’s projection squarely in the upper range of current analyst consensus between $700 billion and $800 billion.
CreditSights revised its own estimates from approximately $600 to $620 billion up to $750 billion. That revision alone represents about 67% year-over-year growth.
The individual company breakdowns tell an even more striking story. Amazon is expected to lead the pack with around $200 billion in single-year capital expenditure for 2026. Alphabet’s capex guidance has been revised upward multiple times, landing in a range of $175 billion to $205 billion. Meta projects spending between $115 billion and $145 billion.
What’s driving the surge #
The demand signals are hard to argue with. Alphabet’s cloud backlog has soared past $460 billion, later climbing to $514 billion, suggesting that customers are lining up for AI-powered compute capacity faster than it can be built. Microsoft’s Azure backlog is reportedly constrained not by demand but by power availability, meaning the bottleneck isn’t customers wanting the product but the physical infrastructure to deliver it.
Goldman Sachs has projected that global investment in AI could exceed $1 trillion by 2026. Hyperscaler capex represents the lion’s share of that figure.
Rising component costs are also contributing to the total. Next-generation GPUs, advanced networking equipment, and the power infrastructure required for increasingly dense data centers all carry higher price tags than their predecessors.
The cash flow question #
There’s an important caveat buried in the optimism. AI-related capital expenditure is anticipated to consume nearly 93% of operating cash flow for these major firms by 2026, leaving very little room for other investments, buybacks, or financial flexibility.
The likely consequence is increased reliance on debt, leasing arrangements, and customer prepayments to bridge the gap. Cloud backlogs suggest strong future revenue streams — a $514 billion backlog at Alphabet, for instance, represents years of contracted revenue that hasn’t yet been recognized.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our