AnswersPublished18 sources
PwC Sees $31.6 Trillion AI Data-Centre Investment Boom by 2050
PwC’s central forecast puts global data centre investment at $31.6 trillion through 2050, with accelerated AI adoption pushing the total near $50 trillion. Annual data centre capital expenditure is projected to rise from about $800 billion in 2026 to $1.1 trillion in 2030 and $1.8 trillion in 2050.
Published byEdited with GPT-5.6 LunaImages generated with GPT Image 1.5
Research answer
Key takeaways
- PwC’s central forecast puts global data centre investment at $31.6 trillion through 2050, with accelerated AI adoption pushing the total near $50 trillion.
- Annual data centre capital expenditure is projected to rise from about $800 billion in 2026 to $1.1 trillion in 2030 and $1.8 trillion in 2050.
- The long term spending shift is from buildings to recurring ICT upgrades: equipment could rise from about 70% of data centre capex in 2026 to 93% by 2050.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: What does PwC’s Global Data Centre Outlook, modeled by Oxford Economics across 46 countries, project about global AI-infrastructure investme. Article summary: PwC projects that AI-driven data-centre capital expenditure will become a sustained, multi-decade investment cycle rather than a one-time construction boom. Its central forecast is $31.6 trillion of global investment thr. Topic tags: general, general web, user generated, government, academic. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, wate
PwC’s Global Data Centre Outlook describes AI infrastructure as a multi-decade investment cycle, not a one-time construction surge. Modeling commissioned from Oxford Economics across 46 countries and territories projects $31.6 trillion in cumulative global data-centre capital expenditure through 2050. If AI adoption accelerates beyond the central forecast, investment could approach $50 trillion. 5
7
The headline figure comes with a major qualification: demand alone will not decide which markets capture the growth. Electricity, grid capacity, advanced chips, connectivity, security and predictable policy are likely to be the practical constraints on new capacity. 5
7
How large is the projected investment?
PwC expects annual global data-centre capital expenditure to grow from roughly $800 billion in 2026 to $1.1 trillion in 2030, reaching about $1.8 trillion per year by 2050 under its central scenario. 5
The forecast includes both the physical facilities and the technology installed inside them. That distinction matters because the model anticipates continued spending after a facility is built: servers, GPUs, CPUs, storage and networking equipment must be refreshed as computing requirements and technology change. 5
7
Equipment—not construction—drives the long cycle
The most important structural finding is the changing composition of capex. ICT equipment accounts for about 70% of data-centre investment in 2026, but PwC expects that share to rise to 93% by 2050. The increase reflects recurring upgrades to chips and servers, which are replaced on cycles of roughly four to six years. 5
7
That creates a different infrastructure pattern from a conventional construction boom. Spending does not end when a building opens; it can recur as operators add computing capacity and replace equipment. For AI infrastructure, the long-term market is therefore tied to both new facilities and repeated hardware upgrades.
Where could the money go?
PwC’s baseline regional outlook projects the largest concentration of investment in the United States, followed by Asia Pacific:
- United States: $15.1 trillion
- Asia Pacific: $8.2 trillion, led by China and India
- Europe: $5.6 trillion
- Middle East: $1.1 trillion
- Africa: $255 billion
5
10
The United States’ projected total represents about 48% of the global forecast. Asia Pacific’s position reflects the scale of expected demand in China and India, but the report’s broader conclusion is that demand is only one part of the location equation. Regions also need the physical and regulatory conditions to turn demand into operating data-centre capacity. 5
7
10
What could push the total higher—or lower?
PwC’s scenarios show how sensitive the outlook is to the pace of AI adoption and the availability of critical inputs.
Faster AI adoption
If businesses and consumers adopt AI more rapidly than in the central scenario, cumulative investment could rise to nearly $50 trillion through 2050. This is an upside case, not a guaranteed outcome. 4
5
7
Tighter chip export controls
Restrictions that disrupt advanced-semiconductor supply chains could reduce cumulative investment to roughly $25.5 trillion through 2050. The mechanism is straightforward: if advanced chips are harder to obtain, some planned computing capacity may be delayed, relocated or scaled back. 4
5
Digital sovereignty
Policies requiring data or computing capacity to remain within national borders have a smaller effect on the overall spending total, according to PwC’s outlook. They can nevertheless redistribute investment toward countries with substantial domestic demand but relatively limited existing data-centre capacity. 5
Why electricity is the decisive constraint
Large AI facilities need electricity that is not only available, but affordable, reliable and increasingly compatible with low-carbon objectives. PwC identifies power availability as the decisive factor shaping where AI infrastructure investment flows, alongside grid access, connectivity, cybersecurity, policy certainty, community consent and access to GPUs. 5
7
The power challenge is already visible in wider energy forecasts. The International Energy Agency’s base case projects global data-centre electricity consumption to rise from about 485 terawatt-hours in 2025 to roughly 945–950 TWh in 2030—approximately doubling in five years. 21
29
That demand does not represent AI alone, since data centres also support other digital services. But the IEA identifies AI as the most important driver of the projected increase. 22
26 As a result, the pace of AI infrastructure investment will depend partly on whether electricity generation and grids can expand quickly enough.
The main implication: investment will be geographically uneven
The forecast does not mean every country with AI ambitions will receive a proportional share of the capital. Projects are more likely to concentrate in markets that can combine power and grid capacity with advanced hardware, strong connectivity, security, supportive regulation and local acceptance. 5
7
That makes AI infrastructure both an investment story and a competition over enabling conditions. The central question is not simply how much demand for AI exists in a market, but whether that market can reliably power, connect and operate the computing capacity needed to serve it.
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "PwC Sees $31.6 Trillion AI Data-Centre Investment Boom by 2050"?
PwC’s central forecast puts global data centre investment at $31.6 trillion through 2050, with accelerated AI adoption pushing the total near $50 trillion.
What are the key points to validate first?
PwC’s central forecast puts global data centre investment at $31.6 trillion through 2050, with accelerated AI adoption pushing the total near $50 trillion. Annual data centre capital expenditure is projected to rise from about $800 billion in 2026 to $1.1 trillion in 2030 and $1.8 trillion in 2050.
What should I do next in practice?
The long term spending shift is from buildings to recurring ICT upgrades: equipment could rise from about 70% of data centre capex in 2026 to 93% by 2050.
Sources
businesstimes.com.sgGlobal data centre spending to reach US$31.6 trillion by ...