Bart, co-founder of Media52 and Professor of Journalism oversees IO+, events, and Laio. A journalist at heart, he keeps writing as many stories as possible.
Artificial intelligence is rapidly becoming the defining technology of the 21st century. From chatbots and smart factories to autonomous vehicles and advanced weather forecasting, AI is transforming the way we work, live, and innovate. But behind its growing capabilities lies a less visible force: an enormous and rising appetite for energy. A new report from the International Energy Agency (IEA) reveals the dual role AI will play in the global energy landscape: as both a major driver of electricity demand and a powerful tool for building a cleaner, more efficient energy system.
Watt Matters in AI
Watt Matters in AI is an initiative of Mission 10-X, in collaboration with the University of Groningen, University of Twente, Eindhoven University of Technology, Radboud University, and the Convention Bureau Brainport Eindhoven. IO+ is responsible for marketing, communication, and conference organization.
More information on the Conference website.
The rise of AI and the demand for power
Since the launch of ChatGPT in 2022, AI adoption has surged across industries and households. Large-scale models are now trained and deployed in massive data centres that require vast amounts of electricity. According to the IEA’s new special report, Energy and AI, global electricity demand from data centers is set to more than double by 2030, reaching around 945 terawatt-hours (TWh), more than Japan’s total electricity consumption today.
AI-focused data centres, in particular, are energy-intensive, with the largest facilities under construction expected to consume as much electricity as two million households. The United States is leading this growth, with data centres projected to account for nearly half of the country’s increase in electricity demand through 2030. By then, more electricity will be consumed for data processing than for manufacturing aluminium, steel, cement, and chemicals combined.
Globally, the footprint of AI data centres is still modest - just 1.5% of total electricity use in 2024 - but that share is growing fast. In advanced economies where electricity demand has been flat or declining for years, data centres could drive more than 20% of demand growth this decade.
Watt Matters in AI
Watt Matters in AI is a conference that aims to explore the potential of AI with significantly improved energy efficiency. In the run-up to the conference, IO+ publishes a series of articles that describe the current situation and potential solutions. Tickets to the conference can be found at wattmattersinai.eu.
View Watt Matters in AI SeriesCan the grid keep up?
Meeting this explosion in demand won’t be easy. The IEA warns that electricity grids in many countries are already under pressure. Grid connection queues are long, and building new transmission infrastructure can take years. In fact, the report estimates that one in five planned data centre projects could face delays due to grid bottlenecks.
“The risk is not just that AI could overwhelm our grids,” said IEA Executive Director Fatih Birol. “It’s that we might be forced to choose between electrifying our economies and powering our digital infrastructure. That’s not a trade-off we want to make.”
The IEA urges countries to accelerate investment in electricity grids, improve the efficiency of data centers, and incentivize more flexible operations. For example, encouraging data centres to shift workloads or use onsite energy storage during peak demand could ease grid congestion.
A mixed energy supply, led by renewables
To power this AI-driven future, a broad mix of energy sources will be required. The IEA forecasts that renewables and natural gas will dominate, thanks to their cost-effectiveness and availability. Half of the additional electricity needed for data centers will come from renewable sources, particularly solar and wind, complemented by grid storage and dispatchable sources such as gas and nuclear power.
Interestingly, the tech sector is emerging as a key player in shaping energy investments. Companies like Google, Amazon, and Microsoft are already among the largest corporate buyers of renewable energy and are now investing in newer technologies, including geothermal and small modular nuclear reactors (SMRs). The first SMRs could come online around 2030, providing zero-emission, 24/7 power for data centres.
AI as a tool for energy optimization
While AI requires significant energy, it also offers powerful solutions for enhancing the energy system's efficiency, reliability, and sustainability. Across oil and gas, electricity, industry, transport, and buildings, AI applications are being used to optimise operations, reduce emissions, and cut costs.
In power grids, AI can help forecast renewable energy production, detect faults, and optimize electricity flow, thereby reducing outages and unlocking previously unused capacity. According to the IEA, AI-based grid management tools could free up 175 GW of transmission capacity globally, more than enough to meet projected data centre growth through 2030.
In industry, AI is being used to fine-tune manufacturing processes, improving energy efficiency and productivity. In buildings, smart systems controlled by AI can reduce energy use for heating and cooling, while in transportation, AI helps optimize routes and vehicle maintenance.
Accelerating clean energy innovation
One of the most promising uses of AI lies in energy innovation. By simulating millions of scenarios and materials, AI can dramatically speed up the discovery and commercialization of next-generation clean technologies. The IEA highlights its potential in battery design, solar PV materials, CO₂ capture, and synthetic fuel production.
But there’s a catch: only 2% of equity raised by energy startups goes to AI-enabled ventures. To address this, the IEA recommends policy support to increase funding, enhance access to data, and establish the digital infrastructure necessary for AI-driven R&D.
Managing risks and reducing emissions
The AI boom is not without risks. Data centers require critical minerals like gallium and rare earths, whose supply is heavily concentrated, posing energy security challenges. Meanwhile, AI also increases the threat of cyberattacks, although it can help defend against them as well.
In terms of emissions, the IEA projects that data center-related CO₂ output could rise from 180 million tonnes today to 300 million tonnes by 2035, or up to 500 million tonnes in a high-demand scenario. However, these figures remain below 1.5% of total energy sector emissions, and could be offset by AI-enabled efficiency gains elsewhere.
Still, rebound effects like increased car travel due to autonomous vehicles could undermine some of AI’s climate benefits. As the IEA cautions, “AI is not a silver bullet. It must be guided by smart policy to realise its potential without adding to the problem.”
A call to action
To unlock AI’s full promise for energy while managing its costs, the IEA recommends three pillars of action:
- Accelerate energy infrastructure investment, especially in clean generation and modern grids.
- Boost efficiency and flexibility in data centres and digital systems.
- Deepen collaboration between governments, energy providers, and tech companies.
Countries that lead in this transition, combining digital and energy strategies, will gain a competitive edge in the AI era. But delay could mean gridlock, both literally and figuratively.
As Birol concludes: “AI is a tool, potentially an incredibly powerful one. But how we use it is up to us. We must act now to ensure it becomes a driver of sustainability, not a barrier to it.”