The artificial intelligence boom is quickly becoming an infrastructure boom.
Global investment in AI infrastructure could reach $31.6 trillion through 2050, according to a new PwC forecast. The money will support data centers and the computing equipment packed inside them. That includes GPUs, servers, storage systems, CPUs, and networking hardware.
The number is huge, but it isn’t fixed. PwC’s central forecast sits inside a much wider range. Depending on AI adoption, cumulative investment could land between roughly $22 trillion and $50 trillion.
And this isn’t a one-time construction rush. AI hardware becomes outdated quickly. That means operators will keep replacing expensive equipment long after the data center itself is built.
AI Infrastructure Spending Could Reach $31.6 Trillion
PwC expects global data center capital spending to reach $31.6 trillion through 2050. The forecast was developed using modeling commissioned from Oxford Economics. It covers 46 countries and territories across five regions. Annual spending could rise sharply during that period. PwC expects it to climb from around $800 billion in 2026 to $1.1 trillion in 2030. By 2050, annual investment could reach $1.8 trillion.
The Real Cost Isn’t Just Building Data Centers
Buildings are only part of the AI infrastructure bill. Much of the money will go toward equipment inside those facilities. AI workloads depend heavily on GPUs and other accelerators. Servers, storage, networking equipment, and CPUs also require regular upgrades. PwC expects ICT equipment to rise from 70% of total spending in 2026 to 93% by 2050. That makes hardware replacement one of the biggest long-term costs.
AI Hardware Creates a Constant Upgrade Cycle
A conventional building can operate for decades with periodic improvements. AI computing hardware has a much shorter life cycle. PwC expects GPUs and servers to require replacement roughly every four to six years. A data center could therefore experience several major technology refreshes during its lifetime. Faster AI innovation could make that cycle even more expensive. New models often demand more powerful chips, memory, networking, and cooling systems.
Faster AI Adoption Could Push Spending Toward $50 Trillion
The $31.6 trillion figure represents PwC’s central scenario. The eventual total will depend heavily on how quickly AI adoption grows. Faster adoption could push cumulative spending close to $50 trillion. Slower growth would bring the figure down significantly. Even then, PwC expects infrastructure spending to remain elevated. Businesses are already moving AI into software, cloud services, industrial systems, and consumer products. Each new workload adds pressure for more computing capacity.
The United States Could Capture Nearly Half of Global Investment
The United States is positioned to take the largest share of the AI infrastructure boom. PwC projects about $15.1 trillion in cumulative U.S. investment through 2050. That represents roughly 48% of the global central forecast. The Americas overall could attract $16.5 trillion. The U.S. benefits from its advanced chip ecosystem and major hyperscalers. Many leading AI developers are also based there.
Asia-Pacific Could Attract $8.2 Trillion
Asia-Pacific is expected to become another major destination for AI infrastructure capital. PwC forecasts about $8.2 trillion in cumulative investment through 2050. China and India are expected to generate significant demand. Both have huge populations and expanding digital economies. AI adoption across businesses and consumer services could increase computing requirements further. Other Asian economies may also benefit as data sovereignty becomes more important.
Europe Faces Power and Planning Constraints
Europe could receive about $5.6 trillion in cumulative investment through 2050. That is below its share of the global economy. Power constraints are one reason for the gap. Planning challenges and fragmented regulation could also slow new projects. Nordic countries may have an advantage. Their renewable-heavy grids and cooler climates can reduce operating and cooling costs. PwC sees energy availability as a major factor in future investment decisions.
The Middle East Is Emerging as an AI Infrastructure Market
PwC expects the Middle East to attract about $1.1 trillion in cumulative investment through 2050. Several countries are already building sovereign AI capabilities and large data center projects. Saudi Arabia and the UAE have become particularly active in this area. However, access to advanced chips remains a potential risk. Trade restrictions could change where future infrastructure projects are built. The region’s ability to secure computing hardware will therefore matter.
Africa Could Receive $255 Billion in Investment
Africa receives a smaller share under PwC’s central forecast. The continent could attract around $255 billion through 2050. However, data sovereignty could change that picture. More governments may want sensitive data stored within their own borders. That could encourage additional local data center construction. PwC’s sovereignty scenario increases Africa’s cumulative investment to around $284 billion.
Electricity Could Become AI’s Biggest Infrastructure Bottleneck
Money may not be the hardest part of the AI buildout. Electricity could be the bigger constraint. AI data centers require enormous amounts of reliable power. They also need grid connections that can support dense computing systems. PwC identifies power as the leading factor shaping future investment locations. Transmission capacity and substation availability could delay projects. Transformer lead times are another concern. Markets with affordable and reliable electricity may gain an important advantage.
Chip Restrictions Could Cut Trillions From the AI Buildout
Geopolitics could dramatically change the investment forecast. PwC modeled a scenario involving tighter restrictions on advanced chip supplies. Under that scenario, cumulative global spending falls to about $25.5 trillion through 2050. That is roughly $6 trillion below the central forecast. The disruption would be strongest during the early years. Spending could fall to around half the central projection by 2030 before supply chains adapt.
Sovereign AI Could Redirect Where the Money Goes
Another major force is sovereign AI. Governments increasingly want critical AI systems hosted on trusted or domestic infrastructure. That could redistribute investment rather than dramatically reduce it. PwC estimates global spending at around $29.5 trillion under its stronger sovereignty scenario. Emerging markets could gain from the shift. India, Vietnam, Indonesia, the Philippines, and Thailand are among those positioned for additional investment. Their domestic demand could support more locally hosted computing capacity.
AI Is Creating a Different Kind of Infrastructure Boom
Railways, electricity networks, and the internet required enormous upfront investment. AI infrastructure has a different financial rhythm. Data centers can remain standing while the technology inside them repeatedly changes. Every new generation of GPUs can trigger another spending cycle. Cooling and power systems may need upgrades as computing density rises.
That makes the $31.6 trillion forecast more than a construction story. AI is creating a recurring market for chips, servers, electricity, cooling, networking, and financing. The countries that can provide those pieces may capture a large share of the next infrastructure cycle. Those that cannot may find the AI boom happening somewhere else.
Sources
Fox Business — AI infrastructure investment projected to top $31T by 2050
https://www.foxbusiness.com/economy/ai-infrastructure-investment-projected-top-31t-2050
PwC — Global investment in AI infrastructure to hit US$31.6 trillion through 2050
https://www.pwc.com/gx/en/news-room/press-releases/2026/global-investment-in-ai-infrastructure.html
