Gartner Forecasts 26% Surge in Data Center Power Use for 2026
AI-optimized servers to consume one-third of total data center electricity
Gartner reports that global data center electricity consumption is projected to reach 565 TWh in 2026, driven by a massive expansion of AI-optimized infrastructure. This surge represents a critical shift in global power demand as specialized AI workloads begin to outpace traditional computing needs.
Key details
Global power consumption for data centers is expected to grow 26% from 447 TWh in 2025 to 565 TWh in 2026. Total worldwide demand is projected to reach 133 gigawatts (GW) in 2026 and escalate to 291 GW by 2030.
AI-optimized servers are the primary driver of this growth; their power consumption is projected to rise 84.2% to 175 TWh in 2026. Gartner estimates that by 2027, AI-specific servers will consume more electricity than conventional servers globally. In the United States, dedicated AI data centers are expected to consume 68 TWh in 2026, accounting for one-third of the nation's total data center power usage.
Why this matters
This forecast underscores the growing "power wall" facing AI development. As compute-intensive GenAI workloads scale, electricity availability is becoming the primary constraint on expansion, shifting the industry focus toward power security and infrastructure efficiency over raw compute capability.
Context
While AI infrastructure is seeing near-vertical demand, traditional data center growth remains relatively flat, with only a 1.2% projected increase for conventional servers in 2026. This stark divergence is forcing grid operators and developers to prioritize high-efficiency cooling and secure access to massive grid capacity years in advance to avoid scaling bottlenecks.
Risks and open questions
Current grid supply may be insufficient to meet the projected demand of 1,200 TWh by 2030. This infrastructure deficit threatens to delay construction projects and inflate operational costs, potentially leading to a competitive "battleground" for the limited remaining grid access in key markets.
What happens next
Infrastructure and operations leaders are expected to pivot toward aggressive efficiency upgrades, including high-efficiency cooling systems and edge computing. Securing direct grid access and exploring on-site power generation will likely become standard requirements for large-scale AI deployments to ensure sustainable growth.
Source: Network World Published on AI Usage Global, author: AUG Bot
