Data Center Capex to Exceed $3 Trillion by 2030 on AI Spending

Infrastructure Investment Accelerates

Global data center capital expenditure is on track to exceed $3 trillion by 2030, propelled by sustained investment in AI infrastructure, according to new research from Dell'Oro Group. The forecast represents a near-doubling from the firm's January 2026 projections, reflecting increased hyperscale spending guidance, expanded global power capacity estimates, and rising commodity costs.

High-end accelerators powering AI-optimized servers will account for the largest portion of data center capex and remain the primary growth driver through the end of the decade, the telecommunications and data center market research firm reported in August 2026.

Why it matters

The dramatic upward revision signals that AI infrastructure buildout is not plateauing but entering a more capital-intensive phase. For technology vendors, cloud providers, and power infrastructure companies, this represents a multi-trillion-dollar market opportunity—but one constrained by power availability and uncertain enterprise ROI. The concentration of spending among the top four U.S. hyperscalers, projected to represent roughly half of global capex, also highlights competitive dynamics that could reshape the cloud market.

Hyperscalers Dominate Spending

The top four U.S. hyperscalers alone are expected to account for approximately half of worldwide data center capex over the forecast period. Meanwhile, a newly tracked segment comprising AI model builders and neocloud service providers is projected to grow at nearly 60 percent compound annual growth rate, outpacing all other customer categories.

"Our 2030 data center capex outlook has nearly doubled since the January 2026 forecast, reflecting higher hyperscale capex guidance, increased projections for global data center power capacity, and higher commodity costs," said Baron Fung, Vice President of Research at Dell'Oro Group.

Beyond Accelerated Computing

While AI accelerators drive headline spending, general-purpose server demand is expected to benefit from expanding inference workloads, agentic AI applications, and storage requirements. This broader infrastructure growth extends beyond pure accelerated computing, according to the research.

However, Fung noted that the pace of expansion will depend on several factors: the sustainability of current investment levels, power availability to support new facilities, and supply chain conditions. Enterprise investment, in particular, remains constrained by uncertain returns on AI deployments.

Efficiency as a Constraint Response

Accelerated and heterogeneous computing architectures, combined with innovations in server efficiency, could help address the escalating cost and infrastructure demands associated with AI workloads. These technological advances may prove critical as power constraints and capital intensity test the limits of current expansion trajectories.

The findings were first reported by Dell'Oro Group in their Data Center IT Capex 5-Year July 2026 Forecast Report, which tracks spending across hyperscale, cloud, colocation, telecommunications, and enterprise customer segments.