The United Nations Economic Commission for Europe (UNECE) warned this week that global data center electricity consumption — driven mainly by AI — will climb from roughly 485 TWh in 2025 to 950 TWh by 2030, close to 3% of world electricity demand, while power grid infrastructure is advancing at a much slower pace.
The core problem UNECE points to isn't just the volume of consumption — it's the mismatch in timelines: a data center can be built and connected to the grid in two to five years, while expanding transmission lines and the rest of grid infrastructure can take over a decade. In several jurisdictions, that gap is already visible in practice: Ireland introduced restrictions on new connections in Dublin, the Netherlands tightened planning decisions due to infrastructure and land constraints, and in the United States, large loads — particularly data centers — are projected to put significant pressure on equipment demand and supply chains through 2030.
The body is calling for greater data transparency, standardized reporting, and real-time information sharing to support planning and operation of power systems, along with closer coordination between data center developers, grid operators, regulators, local authorities, and the agencies responsible for water, land use, and infrastructure. In other words: treat electrical capacity as an input that needs to be planned as far in advance as the compute hardware itself.
Carlos Montiel is an enterprise AI solutions architect. He implements LLMs, agents, RAG, and orchestrators for companies across Guatemala and Latin America. Get in touch for a consultation.
Contact Carlos Montiel