By Martin Ekpeke
Annual global expenditure on data centre infrastructure is projected to reach $1.8 trillion per year by 2050, up from roughly $800 billion in 2026, according to a major report published by PwC.
The Global Data Centre Outlook, the first long-range analysis to project capital expenditure in the sector through mid-century, forecasts a cumulative $31.6 trillion investment pipeline to build and sustain the compute capacity required for advanced artificial intelligence.
Unlike historical infrastructure investments like roads or utilities, where capital spend drops sharply after initial construction, the AI buildout will require continuous, compounding reinvestment.
PwC projects that recurring chip and information technology upgrades will devour the vast majority of long-term capital allocation. ICT equipment, which accounts for 70% of data centre capex today, is expected to grow to 93% of total spending by 2050 as microprocessors reach end-of-life cycles every few years.
“Cumulative spend on data centre infrastructure is projected to reach US$31.6 trillion between 2026 and 2050. Annual spend is expected to grow from approximately US$800 billion in 2026 to US$1.8 trillion by 2050. By 2050, hardware upgrades will account for a 93% share of total capital expenditure,” the report noted.
According to the report, geography will play a massive role in where this vast pool of capital settles, creating clear regional concentrations while facing distinct power bottlenecks. The United States is expected to capture 48% of cumulative global investment, totaling $15.1 trillion, anchored by its strong position in the advanced-chip ecosystem. Meanwhile, the Asia-Pacific region is forecast to absorb $8.2 trillion, driven primarily by expanding markets in China and India. In Europe and the Middle East, investment is being accelerated by government-backed sovereign AI strategies seeking localized compute power.
However, power capacity remains the primary constraint. PwC highlights access to scalable, reliable, and low-carbon electricity as the single most critical factor dictating where future facilities can realistically be built, followed by connectivity, policy certainty, security, and GPU access.
“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,” Clara Cutajar, Global Infrastructure Leader at PwC Australia explained.
The 46-country study models two critical macroeconomic risks that could reshape global compute capacity.
First, stricter trade barriers and chip export controls could cause annual investment to plummet to half of baseline projections by 2030, shrinking overall cumulative investment through 2050 by nearly $6 trillion down to $25.5 trillion even with supply chain adaptation.
Second, if nations focus heavily on localized, sovereign data infrastructure, total global spend would drop only marginally, but capital would shift away from centralized tech hubs toward emerging domestic markets where governments enforce local data control.