30 companies shaping
the AI infrastructure race.

Designed to scale toward 5 GW. This is the maximum planned cluster scale, not today’s energized IT load.

The full campus figure includes the active Trainium AI cluster and continued expansion; AWS does not publish a separate live IT-load figure.

Microsoft says roughly 2 GW will be delivered over five to seven years, so the ranking uses the final planned scale.

The site can support up to 1.6 GW. Its initial 700 MW remains subject to phased utility studies and tenant delivery.

The first investment phase is one part of a 1.5 GW renewable-powered AI factory planned to enter service from 2028.

The full-build plan spans ten data centers engineered for dense, liquid-cooled AI deployments.

The up-to figure comes from a Microsoft Vera Rubin compute letter of intent, not currently energized capacity.

The dedicated cluster starts with a 200 MW phase inside the wider 5 GW UAE–US AI campus.

Google says “gigawatt-scale” but gives no exact MW. A 1,000 MW proxy is used only to place the project in this ranking.

Reliance describes the wider campus program as gigawatt-scale. The separately contracted first Meta facility is 168 MW.

A future cloud and AI campus whose published figure represents the completed site rather than an operating hall.

NEXTDC markets S7 as a 550+ MW AI-ready campus; delivery remains subject to approvals and phased construction.

Published campus capacity across six cloud and AI-ready facilities, not one data hall.

The operational NVIDIA AI deployment occupies only part of the wider renewable-powered data center park.

The 500 MW campus begins with an initial 250 MW development phase.

The figure is covered by heads of terms; final partnership terms and phase delivery still matter.

The final campus is planned for 480 MW; its first construction phase accounts for 240 MW.

The company filing reports 100 MW energized with the remaining campus capacity under construction in 2026.

The long-range IT-capacity pipeline spans several buildings and delivery phases rather than one live hall.

The operator publishes campus scale; tenant-specific AI load remains undisclosed.

The live hyperscale campus supports high-density AI deployments but is not used exclusively for AI.

The initial committed facility is 100 MW, with expansion potential taking the site toward 300 MW.

The published total covers three AI-ready buildings at different delivery stages.

The figure covers the entire expanding campus; each individual operational building is smaller.

A hyperscale campus designed for cloud and AI/ML workloads, with capacity delivered in phases.

The AI-ready campus is under construction, with its first phase expected to enter service in 2028.

A purpose-built AI-ready hyperscale campus whose published capacity represents full development.

The total IT capacity spans three AI-ready facilities; the first delivery is planned for 2027.

A planned London-region facility; the figure is not current operating load.

The substation is live while the data halls remain under phased construction.