Europe urged to prioritise proven AI data centres for grid access

Why grid access is under scrutiny

The International Energy Agency projects global data centre electricity consumption will reach 945 terawatt-hours by 2030, more than double the 2024 level. This surge is driven largely by AI workloads, which concentrate demand in specific locations and strain local grids.

An MIT study found 95% of generative AI pilots failed to deliver measurable returns, raising doubts about the economic viability of many proposed data centres. Without proven profitability, regulators face opportunity costs in allocating finite grid capacity to projects that may not materialise.

In Ohio, AEP Ohio required financial commitments from data centre developers before granting grid access. Initial requests for 30 GW of capacity dropped to 13 GW after engineering study payments were mandated, and only 5.6 GW secured binding agreements.

Infrastructure delays and opportunity costs

Data centres can be built in two to three years, but the energy infrastructure they rely on—such as natural gas turbines and transformers—takes longer. Lead times for new turbines extend up to five years, creating a mismatch between demand and supply.

Allocating grid capacity to AI data centres may divert resources from industries like manufacturing, defence production, and cloud services. Regulators are advised to weigh these trade-offs before approving requests.

Calls for stricter approval criteria

According to The Next Web, Europe should "resist treating data centres and AI power demand as an automatic priority" and instead require developers to demonstrate credible milestones, financial commitments, and transparency in water usage.

The proposal suggests developers cover grid upgrade costs attributable to their projects. The argument is that AI should "earn its place in the energy system through credible commitments and demonstrable value," echoing the approach taken by AEP Ohio.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

For AI startups and operators, this signals a shift toward higher scrutiny and accountability in energy-intensive projects. Securing grid capacity may soon require upfront proof of demand, binding financial commitments, and transparency on resource use—barriers that could slow down speculative projects while favouring those with clear business models and public benefits.

For investors, the focus on profitability and infrastructure costs could reshape due diligence, particularly for early-stage AI infrastructure plays. The Ohio case shows that even large-scale demand projections can evaporate when developers are asked to put skin in the game.

The debate also highlights an emerging tension between AI’s growth ambitions and the practical constraints of energy systems. Startups that can align their data centre strategies with local energy policies—such as co-locating with renewable sources or demonstrating efficiency gains—may gain a competitive edge.

Related on Lazyfounder

• More ai news
• MIT