A source-linked institutional snapshot of the capital build-out behind artificial intelligence: hyperscaler capex, data-center investment, compute demand and the power constraint shaping where new capacity can be built.
PwC baseline projection for cumulative global AI-infrastructure capital expenditure through 2050.
PwC estimate for annual data-center capital expenditure in 2026.
TrendForce estimate for combined 2026 capex of nine major cloud service providers.
MIDA reported cumulative data-center-related investment from 2021 through H1 2026.
The most useful 2026 signal is not a single company budget. It is the convergence of hyperscaler spending, data-center construction, semiconductor upgrades and power availability. These figures are not directly additive because they use different scopes and accounting bases; the tracker keeps each source in its original frame.
PwC projects cumulative global AI-infrastructure investment of US$31.6 trillion through 2050, with annual data-center capex rising from roughly US$800 billion in 2026 to US$1.8 trillion by 2050.
TrendForce estimated that combined 2026 capex for Google, Amazon, Meta, Microsoft, Oracle, ByteDance, Tencent, Alibaba and Baidu would exceed US$886.7 billion, approximately 90% higher year over year.
MIDA reported RM95.8 billion of data-center and cloud-computing projects in the referenced 2026 period, while cumulative data-center-related investment from 2021 through H1 2026 reached RM385.7 billion.
NEXUS preserves the stated scope of each source rather than normalizing incompatible figures into a single artificial total.
Researchers, journalists and analysts may reference the figures on this page with attribution to the underlying primary source and NEXUS PROJECT as the compilation layer. For institutional use, verify the linked source before relying on a figure in a transaction or investment decision.