Global data centres are set to emit 70% more carbon emissions by 2030, according to a stark new warning from the International Energy Agency (IEA). The Paris-based energy watchdog says the breakneck expansion of artificial intelligence is driving a surge in electricity demand that will likely be met with fossil fuels in the US and China.
The IEA’s influential annual report found that data centres will consume twice as much electricity by 2030 compared with today. AI accounts for three-quarters of this growth, with AI data centres alone expected to consume an extra 350 terawatt-hours (TWh) of electricity every year within four years—more than the annual electricity consumption of Great Britain.
Why Data Centre Emissions Are Set to Soar
The projected 70% rise in carbon emissions from data centres is not simply a result of more servers. It reflects the energy mix powering those servers. As the US and China race to lead in AI, both countries are turning to fossil fuels to meet the enormous electricity needs of training and running AI models.
The IEA’s executive director, Fatih Birol, said the energy world is being reshaped by policy decisions made in response to recent crises. “I don’t see any credible pathway for the energy world to return to the way it was before,” he said. The report highlights that the Middle East war has caused the greatest energy market disruption in history, adding uncertainty to the global energy future.
Data centre developers have emerged as the world’s biggest corporate buyers of renewable energy. However, the sheer scale of AI’s electricity demand means that renewable additions alone are not keeping pace, leading to increased reliance on fossil fuels.

The AI Energy Dilemma: Growth vs. Sustainability
The IEA’s findings underscore a fundamental tension: AI is both a potential solution to climate change and a growing source of emissions. While AI could help optimise energy grids and reduce emissions, its own energy footprint is expanding rapidly.
This dilemma is particularly acute in the US and China, where AI investment is surging. The IEA notes that the rising need for air-conditioning and a growing global population are also driving up energy demand, further complicating efforts to decarbonise.
For the tech industry, the report is a call to action. Companies must balance their AI ambitions with investments in clean energy and efficiency. Otherwise, the carbon footprint of the digital economy will continue to grow unchecked.
Key Takeaways from the IEA Report
- 70% emissions increase: Data centres will emit 70% more carbon by 2030 compared to current levels.
- Double electricity consumption: Global data centres will use twice as much electricity by 2030.
- AI drives three-quarters of growth: AI accounts for 75% of the increase in data centre electricity demand.
- 350 TWh extra annually: AI data centres will consume an additional 350 TWh per year—more than the UK’s total annual electricity use.
- Fossil fuel reliance: The US and China are turning to fossil fuels to power AI growth, undermining climate goals.
What This Means for the Energy Sector
The IEA’s warning has significant implications for energy utilities and policymakers. Utilities face mounting pressure to meet soaring demand while decarbonising. In the US, natural gas is increasingly seen as a stopgap, while China continues to rely heavily on coal.
This trend could lock in decades of carbon emissions and make it harder to meet global climate targets. It also raises questions about the sustainability of the AI boom. As data centre emissions rise, so too will scrutiny from regulators and investors.
Some relief may come from AI itself. A separate study suggests that AI could cut global emissions by 5-10% by 2030 through efficiency gains and optimisation. But that potential is far from guaranteed, and it may be outweighed by AI’s own energy consumption.
Comparing Energy Demand: Data Centres vs. Countries
To put the scale of data centre electricity demand into perspective, consider the following comparison. The additional 350 TWh per year that AI data centres will require is roughly equivalent to the entire annual electricity consumption of a country like Great Britain. The background to this story is set out in Carbon Emissions From Fossil Fuels Fall Slightly as Iran War Boosts. A related thread runs through Private Equity Firms' Energy Assets Emit 1.5 Billion Tons.
| Metric | Data Centres (2030 projection) | Comparison |
|---|---|---|
| Additional electricity demand from AI | 350 TWh per year | More than UK’s annual electricity consumption |
| Total electricity consumption growth | 2x current levels | Equivalent to adding a country the size of Japan |
| Carbon emissions increase | 70% by 2030 | Compared to today’s levels |
| AI share of data centre demand growth | 75% | Three-quarters of total growth |
The table illustrates the enormity of the challenge. Meeting this demand without a corresponding rise in emissions will require massive investment in clean energy and grid infrastructure. Earlier coverage in US Datacenter Emissions Equal Denmark's Annual Carbon Output shows how the argument developed. A closer look at the same subject appears in How to Tackle Data Centre Climate Impact and Carbon Emissions.
FAQ
Why will data centre emissions rise by 70% by 2030?
The rise is driven by the rapid expansion of AI, which is causing data centres to consume twice as much electricity by 2030. Much of this additional demand will be met by fossil fuels, particularly in the US and China, leading to a 70% increase in carbon emissions.
How much extra electricity will AI data centres consume?
AI data centres will consume an additional 350 terawatt-hours of electricity per year by 2030. This is more than the total annual electricity consumption of Great Britain.
What did the IEA say about the future of energy?
The IEA’s executive director, Fatih Birol, said there is no credible pathway for the energy world to return to pre-crisis conditions. The future will be shaped by policy decisions made in response to the Middle East war and the growing demand for energy.
Are data centres buying renewable energy?
Yes, data centre developers have become the world’s biggest corporate buyers of renewable energy. However, the scale of AI’s electricity demand is so large that renewables alone are not sufficient, leading to increased reliance on fossil fuels.
The IEA’s report makes clear that the energy landscape is at a turning point. As data centre emissions climb, the pressure is on governments and tech companies to reconcile AI’s promise with its environmental cost. The decisions made in the next few years will determine whether the digital revolution accelerates or mitigates the climate crisis.
For a closer look at how UK data centres are already on track to outpace even ExxonMobil in emissions, see our analysis of UK datacentres to emit more carbon than ExxonMobil. That piece breaks down the country-level impact of the AI power grab.
On the other side of the ledger, there is evidence that AI itself could help reduce emissions. A recent study suggests that AI could cut global emissions 5-10% by 2030 through smarter energy management. The net effect will depend on how quickly the tech sector can decouple its growth from fossil fuels.
Utilities and grid operators are already feeling the strain. The IEA’s findings suggest that without significant policy intervention, the surge in data centre electricity demand will lock in higher emissions for decades. Investors and regulators are beginning to take notice, and the coming years will likely see tougher scrutiny of AI’s energy footprint.
Ultimately, the IEA’s warning is a reminder that the energy transition is not automatic. It requires deliberate choices. As UK datacentres to emit more carbon than ExxonMobil shows, the stakes are high. And as AI could cut global emissions 5-10% by 2030 suggests, there is still room for optimism—if the right investments are made.