Hyperscalers to spend over $805 billion on AI in 2026

A shift in capital allocation by the world's largest technology companies is set to prop up the U.S. economy, as hyperscalers like Meta Platforms (NASDAQ: META) and Alphabet (NASDAQ: GOOGL) divert funds from shareholder returns to an unprecedented arms race for artificial intelligence infrastructure. The five largest hyperscalers are expected to spend approximately $805 billion on capital expenditures in 2026 alone, a structural trend that is replacing share buybacks as the primary use of cash.

"Everything we build with is sold," Nebius Group (NASDAQ: NBIS) CEO Arkady Volozh said on a recent earnings call, a sentiment that captures the insatiable demand for AI computing power. The full-stack AI cloud company raised its midpoint capital spending guidance to $22.5 billion after reporting 684% year-over-year revenue growth in the first quarter.

The spending figures are staggering. Meta has increased its 2026 capex forecast to a range of $125 billion to $145 billion, while Alphabet is targeting $180 billion to $190 billion. This firehose of investment comes at a direct cost to shareholders. After spending a combined $27.9 billion on share repurchases in the first quarter of 2025, the two tech giants spent nothing on buybacks in the first quarter of 2026, according to company filings.

This pivot from shareholder returns to infrastructure investment is now the primary support pillar for the U.S. economy, according to Morgan Stanley. The bank forecasts nonresidential business fixed investment to grow 7.0% in 2026 and 8.0% in 2027, a stark contrast to the 1.8% growth projected for real consumer spending this year. This spending is more structural than cyclical, making it insensitive to the oil shocks or consumer sentiment that typically dictate investment cycles.

The Meta deal changes everything

The scale of these long-term commitments is exemplified by Nebius Group's recent contract with Meta. The five-year deal, valued at $27 billion, includes a $12 billion commitment for dedicated compute capacity starting in 2027 and a $15 billion option that gives Nebius flexibility to sell capacity to Meta or capture higher prices on the open market.

The deal provides Nebius, which only recently relaunched after spinning off from Russian internet company Yandex, with a financial backstop from a top-tier customer. It also highlights the competitive pressure on hyperscalers to secure computing power years in advance. Nebius, in turn, has deepened its partnership with Nvidia (NASDAQ: NVDA), securing early access to next-generation hardware like the Vera Rubin GPU platform.

Capex over consumption

The trend has led Morgan Stanley's Chief U.S. Economist Michael Gapen to frame the mid-year economic outlook as "Capex Over Consumption." While the bank forecasts U.S. real GDP growth of 2.3% in 2026, the foundation is uneven. Higher energy costs are neutralizing fiscal benefits for households, causing consumer spending to decelerate.

Against this weakening consumer backdrop, corporate AI spending is the main driver. The investment is not just about building data centers; it's a full-stack arms race. Nebius's 2026 acquisitions of Eigen AI for inference speed, Clarifai for system optimization, and Tavily for agentic search show how companies are buying their way up the AI stack. For investors, the dynamic creates a clear trade-off: the near-term tailwind of share buybacks has been sacrificed for a long-term, high-stakes bet on AI dominance.

This article is for informational purposes only and does not constitute investment advice.