Land. Power. Compute.: CUDO brings an industry together around AI delivery

CUDO Compute’s Land. Power. Compute. report examines a defining challenge for AI infrastructure: bringing the physical components of deployment together at the right place and time. Its argument is that access to processors alone cannot deliver scale. Suitable land, reliable power, cooling, permissions and computing systems form an interconnected operating environment.

The report brings together perspectives across power, engineering, construction, colocation, legal services, technology and delivery. CUDO acknowledged contributors and partners including Ark Data Centres, AVK-SEG, blocz, Cambridge Management Consulting, Conapto, Canopy, CloudAdvise, ClusterPower, Digital Realty, Holmes & Hills Solicitors, Intel, Kao Data, Novara Infrastructure, Schneider Electric, Supermicro and WSP.

That breadth matters because infrastructure decisions affect one another. A site’s electricity connection influences its delivery timetable. Cooling requirements affect building design. Hardware choices interact with power density and the capacity of the surrounding systems. The report presents these as connected decisions, rather than separate procurement exercises.

CUDO also examines the mismatch between fast-moving computing technology and the longer development cycles of grids, sites and approvals. Its emphasis is on infrastructure that can support successive hardware generations, with available power and deliverable sites shaping the plan from the outset.

For Bolt, the report provides a clear expression of the operating approach behind its interest in CUDO. The commercial value of infrastructure lies in the service it enables: capacity that customers can access reliably, within a workable delivery schedule and cost structure. Coordination across disciplines is central to that outcome.

Land. Power. Compute. gives that wider industry conversation a practical focus. By bringing contributors from across the delivery chain into one report, CUDO connects the ambition surrounding AI with the organisations responsible for making deployment possible. The full report explores those dependencies in greater depth, from site selection and energy economics to thermal management and the future of physical AI infrastructure.