Key takeaways

- AI technologies are becoming more commonplace in business and consumer services. This trend is creating challenges, especially as new use cases emerge across a wide range of industries. 

- AI requires large amounts of electricity to run, so meeting the growing electricity demand of greater AI adoption could place a significant strain on electricity systems if not properly managed.

- AI also holds the potential to make electricity system management more efficient and secure, through devices such as smart meters that can crunch data to predict future changes in electricity demand. 

- This AI-powered forecasting enables more precise matching of demand with supply from available electricity generation sources, helping to minimise costs and maintain energy security.

- Meeting the increased electricity demand from AI must not compromise global climate goals, making the development of secure, low carbon electricity systems all the more urgent.  

For some, the phrase artificial intelligence (AI) once conjured images of humanoid robots. As technology has advanced, AI capabilities are now being concentrated more into consumer apps and business services, from chatbots and image recognition to data analysis and automation. AI’s ability to process large amounts of data and swiftly identify previously unseen patterns, means industries as diverse as healthcare, manufacturing and entertainment are implementing the technology at an ever-increasing rate.  

A 2024 survey by McKinsey found that 72% of companies globally are adopting AI technologies, driven in part by the launch of new Generative AI programmes like OpenAI’s ChatGPT and Google’s Gemini.  

As the technology becomes more widely adopted, the computing power needed to deliver AI services is growing quickly. The computing power used to train notable AI models has doubled every six months since 2010, creating a real increase in electricity demand to power AI. According to the International Energy Agency (IEA), global data centre electricity demand will more than double from 2022 to 2026 with AI playing a significant role in that increase. 

Meeting the growing electricity demand of greater AI adoption presents a real challenge for global energy systems: How can the world power an AI future while continuing to reduce carbon emissions and meet global climate goals? 

Nevertheless, AI also presents an opportunity to help make electricity systems more efficient, sustainable and secure by looking at trends and data to help manage them better. 

“Using AI in energy systems can help predict the electricity demand of businesses and customers and the energy consumption of artificial intelligence,” explains Bjoern Reinke, Director of Data and Data Science. “Those predictions can be combined with weather forecasts to understand what the electricity grid can actually produce and what demand it can cope with.”

Utilizing AI to forecast electricity trends and understand what new infrastructure will be needed to meet future electricity demand starts with data and creating smarter grids.  

Building AI-powered smart grids 

From how many times you boil the kettle per day, to how many loads of washing you do per week, your accumulated electricity usage all translates into data. Smart meters, in homes, businesses, transport networks and cities, can all collect valuable information that grid operators can use to identify patterns and create electricity forecasts. 

“Smart meters can track energy consumption and over time that consumption becomes really predictable. When something becomes predictable, it becomes easier to manage,” says Reinke. “The more individuals and businesses have smart meters, the more reliable the energy consumption forecasts become, which is why the smart metering program continues to be so important.”

For grid operators, this AI-powered forecasting enables more precise matching of demand with supply from available electricity sources. The increased efficiency of AI-enabled smart grids contributes to minimizing costs and maintaining energy security. 

Smart grids can also play a vital role in the global energy transition, as countries move from fewer, large fossil fuel power stations to a more complex mix of decentralized, varied sources of renewable supply and low carbon storage. AI can rapidly bring together disparate data sets from a broad range of electricity generation sources, and balancing services, to ensure demand is met with low-carbon solutions and that systems operate reliably and efficiently while decarbonizing. 

Reducing costs and risk through AI

AI’s potential isn’t limited to matching electricity supply and demand. Another key application of the technologies is within the infrastructure and operations of power systems. 

“The electricity system itself is a very complicated network of expensive assets, including electricity generation and storage, as well as transmission systems,” explains Reinke. “Thereby, the maintenance process and the reliability of the full set of assets is very important to energy security, as well as costs and risk reduction.” 

Machine learning technology is used to continuously monitor and analyze the performance of electricity assets to identify potential faults ahead of time.

“AI can be used in things like temperature and vibration monitoring to detect when any of these assets may be likely to fail or reduce in effectiveness,” adds Reinke. “It’s this predictive maintenance that enables electricity system operators to be more targeted and efficient with grid maintenance budgets – a cost saving that can then be passed on to customers.”

Predictive maintenance and anomaly detection is not only restricted to system-wide deployments. 

“AI technologies are not just used for predictive personal safety but also operational and supply chain safety,” stresses Reinke. “It is this preventative and predictive maintenance enabled by AI that reduces both cost and risk.”

In addition, AI does not just offer opportunities to reduce cost and improve efficiencies but also helps to spur new innovation in renewable, dispatchable power generation and carbon removals. It requires huge demand for concentrated and stable power, providing the opportunity for the next generation of renewable energy technologies.

Bioenergy with Carbon Capture and Storage (BECCS), which is being developed by Elimini, uses a supply chain of sustainably sourced biomass to generate renewable, baseload power, while capturing carbon in the process and storing it permanently and safely in specific geological sites. 

“BECCS at that scale is very uncommon. Industries are still learning and Elimini is one of the pioneers in doing that, which means that we look at the data sets created by the end-to-end BECCS process,” says Reinke. “Optimizing bioenergy emissions and capturing the carbon in those emissions in a really sustainable and efficient way is a real opportunity for AI.”

This AI-enabled optimisation of the BECCS process can also include analysing the sustainability of potential biomass fiber baskets, monitoring transport and supply chain emissions, maintaining power station operations, or prospecting optimum carbon capture sites. 

The technology could even help determine forestry and carbon storage locations where the social and economic environment is well suited for a new BECCS plant.

AI’s potential to transform the energy system and create efficiencies is undeniable, but to harness its capabilities and manage new levels of electricity demand, the technology must evolve sustainably.

Meeting AI’s demand for power while continuing to decarbonize

The proliferation of AI has already created a surge in electricity demand. According to Goldman Sachs research, global AI data centre demand for electricity rose from 8TWh in 2022 to 12TWh in 2023. This demand can be managed, but it must be met by low carbon, renewable sources if companies and countries are going to meet their climate targets. 

Google recently reported that the soaring demands of its AI operations caused a 13% rise in its corporate emissions in 2023 compared to the year before, and a 48% increase since 2019. Renewable power generation will be essential in reversing this trend and meeting AI’s demand for power. 

A diverse range of generation technologies are required to ensure stable, reliable grids that deliver affordable power. This will include intermittent renewables like wind and solar, as well as dispatchable forms such as BECCS, which can deliver 24/7 electricity while removing carbon from the atmosphere.

“Meeting the increased demand for power from AI will require investment in green assets and a clear AI governmental policy,” says Reinke. “At Drax, we believe this policy should be based on transparency of data and collaboration to create a secure and balanced grid for all.” 

While AI will play a significant role in future global economies, this must not come at the cost of compromising on climate targets. The continued rapid decarbonisation of global power systems is needed to ensure AI’s power demand can be met in a sustainable way and help reach global climate goals.