01
Site and power assessment
Sites are compared by available grid capacity, connection timeline, tariffs and climate, before land and equipment are committed.
Energy Infrastructure
Power, storage and cooling for new AI compute capacity.
The growth of AI compute is limited by the availability of power. Aphelium plans the energy infrastructure of new capacity as one system, from the choice of site to operation.
Context
Demand for power grows faster than the grid.
2×
Data center electricity use expected by 2030.
Manba: IEA.
800 MW
New AI compute capacity agreed for Uzbekistan through 2030.
Manba: Government of Uzbekistan, 2026.
Scope
Four areas, planned together.
02
Grid connection and supply
Connection strategy and long-term supply arrangements with grid operators and generators.
03
On-site generation and storage
Solar generation, battery storage and backup power, integrated with the facility load and the grid.
04
Power and cooling for AI data halls
Electrical and cooling design for high-density racks, including liquid cooling, specified for the lowest energy cost over the life of the facility.
Process
From site to operation.
Site
Power, land, climate
Supply
Grid connection, contracts
Generation and storage
Solar, batteries, backup
Facility
Power and cooling design
Operation
Optimization by the Aphelium platform
01One system
Decisions on site, supply, generation and cooling fix the energy cost of a facility for decades. Aphelium plans them together and carries the same physical model into operation, where the Aphelium platform optimizes the facility from the first day.
02Who Aphelium works with
- Data center developers
- Operators expanding capacity
- Investors in compute and energy assets
- State programmes for digital infrastructure