JetherVerse News
AI· 1 min read

IEA: Datacentre Emissions to Rise 70% by 2030

The IEA says datacentre emissions will be 70% higher by 2030, with electricity use doubling and AI driving three-quarters of the growth. AI datacentres alone add 350TWh a year.

By JetherVerse News

The International Energy Agency says emissions from datacentres will be 70% higher by 2030. The Guardian reported the findings on 19 November 2026. Electricity use by datacentres doubles over the period, and AI accounts for three-quarters of the growth.

The numbers are large. AI datacentres will add 350 terawatt hours a year, more than Great Britain uses. The source of the power matters for emissions. About half of China's AI datacentre power will come from coal by 2030. The US will use about 60% gas, with plans for about 180 gigawatts of new gas plants. The EU is expected to use about 70% renewables.

IEA executive director Fatih Birol framed the shift in strong terms. He said he sees no credible pathway for the energy world to return to how things stood before. His point is simple. Datacentre demand locks in new generation, and gas plants built now run for decades.

For technology firms, the report moves energy from a cost line to a strategy issue. Where a datacentre sits decides its carbon footprint. A model trained on a coal-heavy grid produces far more emissions than the same model trained on a renewable grid. Cloud customers increasingly ask providers for region-level carbon data, and some regions already publish this data.

For developers and businesses, there are steps to take now. Pick cloud regions with cleaner grids when latency allows. Cache results and avoid unnecessary AI calls, since each call uses power. Use smaller models for simple tasks. These steps cut both cost and emissions.

Nigerian and African operators face a related problem. Power supply is a daily constraint, and local datacentre projects must plan generation along with computing hardware. The IEA figures show the same trade-off at global scale.

Plan for energy cost in your product budget. Power prices and carbon rules differ by region, and both are rising pressures for any team running AI workloads at scale.

Read the full report at The Guardian.

Source: The Guardian

Get the news in your inbox

The tech stories that matter from JetherVerse News, with the source on every one.

Free. Unsubscribe any time.