Average data center rack power density has risen from 12 kW in 2024 to 26 kW in 2026, more than doubling in two years, driven by AI accelerator deployments, space constraints, and improved thermal options such as direct-to-chip liquid cooling and immersion. Individual GPUs in some configurations reportedly draw up to 700 watts per device, and some facility designs are now targeting as high as 1 MW per rack, though that figure remains an outlier. Higher per-rack loads create cascading demands on power distribution architectures, cooling infrastructure, and building structures, since facilities designed for lighter racks may require structural reinforcement and new floor systems. Operators facing permitting delays or moratoriums on new builds are likely to densify existing space, accelerating these pressures further.
The jump from an average of 6.1 kW per rack in 2016 to 26 kW in 2026 represents a generational shift in facility requirements, forcing operators to replace or retrofit power delivery, cooling, and structural systems that were not designed for these loads. Facilities that cannot adapt risk being unable to host AI workloads, directly affecting their competitive position as demand for high-density compute continues to outpace available capacity.
Data Center Knowledge's coverage of rapidly rising rack power levels signals a structural shift in how data centers must be designed and powered, a high-relevance topic given ongoing AI buildout coverage. This is distinct from already-published stories on cooling demand and grid stress.