The issue
AI growth is making electricity strategy central to data centre investment. The Antalya Artificial Intelligence Pledge, released ahead of COP31, brings that relationship into focus. It promotes greater use of zero- and low-emission energy, efficient infrastructure, integration with electricity grids and environmental reporting. Its commitments are voluntary and create no new obligations; businesses selecting additional commitments can set their own targets and dates.1
The scale of the challenge is substantial. In its 2025 Energy and AI report, the International Energy Agency projected global data centre electricity consumption would more than double to around 945 TWh by 2030. It also estimated that grid constraints could delay around one-fifth of planned projects unless addressed.2
Why should businesses care?
For developers, operators and investors, electricity availability determines when facilities can open and generate revenue. Procurement choices then shape operating margins, exposure to price volatility and progress towards emissions targets.
A low contract price does not capture the full cost of serving a facility continuously. Annual renewable electricity matching can support clean generation while leaving gaps between when electricity is generated and consumed. Those gaps affect both market exposure and the emissions associated with electricity use.3
The opportunity also extends beyond individual facilities. Data centres can support investment in cleaner grids and provide valuable flexibility. Poorly coordinated growth can instead increase electricity costs, constrain other users and lock in emissions.2
Our view is that the Antalya Pledge reinforces the case for assessing cost, reliability and emissions together, from site selection through procurement and financing.
What should businesses be doing?
- Make energy strategy part of site selection.
- Assess connection capacity, delivery dates, local generation, network charges and cooling requirements before committing capital.
- Model phased demand growth and the financial consequences of delayed energisation.
- Identify workloads that can shift without compromising customer commitments.
- Procure a portfolio that fits demand.
- Compare utility supply, power purchase agreements, on-site generation, storage and demand flexibility against the facility’s hourly consumption.
- Assess whether procurement supports additional clean supply and how closely generation matches demand in time and location.
- Annual matching alone does not demonstrate continuous clean electricity supply.3
- Test the full investment case.
- Compare total delivered electricity costs, including network upgrades, balancing, storage and backup.
- Stress-test prices, demand, connection delays and supplier performance.
- Evaluate emissions alongside cash flow and returns, then align contract lengths, capital commitments and financing with the project’s development schedule.
- Make commitments measurable.
- Set clear targets for electricity use, emissions and clean supply, supported by auditable procurement evidence.
- Track water use and efficiency alongside energy performance.
- Agree responsibility for infrastructure costs with utilities, and engage affected communities early.
- Where AI is claimed to deliver climate benefits, assess those benefits separately from its own footprint.
Sources
- COP31 Presidency, Antalya Artificial Intelligence Pledge, October 2026.
- International Energy Agency, Energy and AI — Executive summary, 2025.
- International Energy Agency, Advancing Decarbonisation through Clean Electricity Procurement — Executive summary, 2022.