- PwC projects US$31.6 trillion in capital expenditure through 2050 to build the compute capacity needed for AI
- Annual data centre capital expenditure is forecast to rise from roughly US$800 billion per year in 2026 to $1.8 trillion per year in 2050
- Recurring chip upgrades—not construction—will drive the majority of long-term capital investment
- Power will be the decisive factor shaping where AI infrastructure investment flows
- Disrupted trade flows for chips could cut global investment by nearly 20%, while a growing sovereignty push would redistribute rather than reduce global investment
LONDON, 2 September 2026 – Global investment in AI infrastructure will hit a record US$31.6 trillion through to 2050, according to baseline projections in PwC’s Global Data Centre Outlook.
On an annual basis, data centre capital expenditure is forecast to rise from roughly US$800 billion per year in 2026 to $1.8 trillion per year in 2050.
The US, which is central to the advanced-chip ecosystem, is expected to capture almost half (48%) of this investment, at $15.1 trillion. Asia Pacific is expected to account for $US8.2 trillion cumulative capex, led by China and India, while sovereign AI strategies are accelerating investment in Europe and the Middle East.
Unlike traditional infrastructure booms, which taper off after the initial build out, AI infrastructure investment is expected to accelerate as chips and other ICT equipment require upgrades every few years. ICT equipment will account for an increasing share of investment—from 70% today to 93% by 2050.
Energy access core to the data centre rollout
The Outlook identifies five factors that will direct where investment flows globally. Chief of these is power, as affordable, reliable and low-carbon electricity at scale is the hardest requirement for many markets to deliver. Connectivity, security, policy certainty and community consent, along with GPU access, will also influence where investment lands.
“AI infrastructure is becoming one of the defining capital allocation challenges of the next generation. It cuts across technology, energy, real estate, supply chains, regulation and financing. This changes how infrastructure investors need to think about capital requirements, risk and returns.”
Export controls and digital sovereignty shape the capex map
The analysis also tested two scenarios to assess how changes in trade policy, export controls and digital sovereignty could affect global AI infrastructure investment.
In the first scenario, tighter export controls disrupt global chip supply chains. Annual investment falls to around half the central forecast by 2030 before gradually recovering as supply chains adapt. Even with that recovery, cumulative global investment through 2050 is projected at around US$25.5 trillion—roughly US$6 trillion less than the central forecast of US$31.6 trillion.
The second scenario paints a different picture. Rather than significantly reducing overall investment, greater emphasis on digital sovereignty and trusted domestic infrastructure changes where capital is invested. Total global spending reduces only slightly, but investment shifts toward countries with strong domestic demand and relatively underdeveloped data centre capacity as governments and regulated industries prioritise local infrastructure.
“The AI buildout is not a rising tide that will naturally lift all boats. Capturing this investment requires active positioning. Investors should recognize data centres as hybrid assets with a complicated risk profile.”
The Outlook is the first of its kind to offer long-range capital expenditure forecasts through to 2050. It covers 46 countries and territories, spanning both data centre buildings and the technology they contain.
https://www.pwc.com/gx/en/news-room/press-releases/2026/global-investment-in-ai-infrastructure.html
