AI Infrastructure Outlays Could Reach $31.6 Trillion by 2050, PwC Estimates

A new projection from consulting firm PwC puts cumulative global spending on artificial intelligence infrastructure at as much as $31.6 trillion through 2050, underscoring the scale of the data center buildout now underway across the technology sector.

The long-horizon estimate arrives amid already elevated capital expenditure among the largest cloud operators. On its August earnings call, NVIDIA said the top five U.S. hyperscalers are on pace to spend roughly $800 billion on capital projects this year. NVIDIA, which designs the accelerators at the center of most AI data center deployments, expects that figure to climb to $1.3 trillion by 2027.

Notably, that forecast excludes spending from a fast-growing tier of so-called neocloud providers and AI laboratories, including OpenAI and Anthropic, which are leasing and building compute capacity of their own. Analysts note that total industry outlays could therefore run meaningfully higher than the hyperscaler figures alone suggest.

The PwC report frames this spending as sustained rather than cyclical, an assessment that matters for semiconductor suppliers whose revenues are increasingly tied to data center investment cycles. NVIDIA, which operates through its Compute & Networking and Graphics segments and serves customers across the United States, Taiwan, China, Hong Kong, and Europe, has become the sector’s most closely watched bellwether for AI demand.

Shares of NVIDIA traded at $223.67 recently, down 0.73% from the prior close of $225.31, giving the company a market capitalization of approximately $5.56 trillion. The stock’s moves have become a common proxy for sentiment on AI infrastructure spending broadly.

Still, the projections come with caveats. Multi-decade forecasts of this kind depend on continued growth in AI adoption, power availability, and the willingness of cloud providers and enterprises to keep expanding compute footprints. NVIDIA’s own $1.3 trillion hyperscaler estimate for 2027 reflects company commentary rather than contracted commitments, and future capital plans among the major cloud operators remain subject to change.

For now, the figures illustrate the scale at which the industry is planning: hundreds of billions of dollars annually in the near term, with cumulative outlays over the next 25 years measured in the tens of trillions.

Source: original release

What to watch

• NVIDIA’s next quarterly earnings report and any updates to its hyperscaler capital expenditure forecasts
• Capital spending guidance from the top five U.S. cloud providers
• Investment announcements from neocloud providers and AI labs, which are excluded from current hyperscaler estimates
• Follow-up research from PwC and other firms refining the long-term AI infrastructure outlook