Using AI to solve the power challenges caused by… AI

Artificial Intelligence (AI) is a central part of the UK government’s plan to boost growth across the UK. Through its AI Opportunities Action Plan, the government intends using AI to deliver its wide-ranging Plan for Change, including commitments to make Britain a clean energy superpower by 2030.  Elsewhere AI is being touted as a fix for everything from administrative overload in the Civil Service, to the UK’s one million-plus potholes. Admirable, but not to my mind where AI can add the most benefit. 

Much has been made of the increased energy demands of data centres processing AI workloads, as opposed to regular facilities more focused on data storage. In April, the International Energy Agency published a new report, projecting that data centre consumption will more than double by 2030. Yet within the energy sector, AI can be used to cure the very problems it’s been accused of causing in terms of its power demands. 

The challenge: A tremendous demand for energy

AI is here to stay. Not only does the government believe it can fix many of the UK’s issues – potholes aside – it considers AI to be the next frontier. Data centres have been elevated to critical national infrastructure status, local authority decisions preventing data centres from being built have been overruled, and dedicated AI growth zones have been created. 

However, the energy demand of AI is notoriously high, and given its potential to revolutionise parts of our lives in ways we don’t yet know, the rate of increase is expected to be steep too. According to the International Energy Agency (IEA), a single AI-focused data centre may require as much power as 100,000 households. Globally, within the next two years, the AI industry could use as much energy as a country the size of Japan. 

The solution: Using AI to cure itself

Ironically, because of its capabilities, we can use AI to provide the solution for its own increased demand for energy. Across the energy sector, AI is being used to optimise generation, transmission, distribution and consumption. It’s also a major instrument in decarbonising the sector and making net zero a reality.  

AI thrives on data. The energy sector both generates and consumes enormous volumes of it. The information generated by smart meters, remote monitoring sensors, electric vehicle charging and other digital assets feeds into AI algorithms to empower smart grids and actively managed networks that benefit the electricity industry on multiple levels. 

Energy companies use AI to connect, optimise, and control energy assets, such as electric vehicles (EVs), heat pumps, and HVAC (heating, ventilation and air conditioning) systems. It allows suppliers to balance and shift loads in real time by incentivising changes to consumer behaviour.

These demand-side response (DSR) programmes reward people for adjusting when they use power to help balance the load on the grid. Platforms that enable this, such as Octopus Energy’s KrakenFlex, use AI to determine what capacity is needed from DSR programmes, when to call a DSR event, and what incentive to offer. By allowing loads to be shifted and re-shaped, AI also enables electricity providers to create new energy products and tariffs – generating income for future investment in the networks. 

AI is being used to manage both generation and demand for commercial and industrial developments. AI helps optimise the use of distributed energy resources (DERs), like batteries, solar and wind, to meet the peaks and troughs of grid demand. A vast amount of data from these assets, combined with weather forecasts and other key variables, are processed to predict and respond to the variability in energy supply and demand. In short, AI enables these resources to be managed more effectively so they perform better and are discharged or curtailed at the right times.

With AI powering optimisation, asset owners will be able to increase the value from their assets by intelligently using them at times that maximise their financial benefits, i.e., when the market conditions are right. In turn, this will potentially provide a boost to investment in new renewable assets.

Self-sustaining innovation

An important additional benefit from the use of AI to optimise networks and assets is that it helps to ensure new generation comes from renewable sources. Variable renewable energy (VRE) sources are inherently intermittent due to weather conditions changing their output. By actively managing these assets and the network, we can compensate for their intermittency through the use of diverse assets with different output profiles that can ramp up or down in response to the changing conditions. AI algorithms that can respond to voltage fluctuations in milliseconds will aid grid stability, enabling real-time load balancing and power-flow optimisation to reduce transmission losses.

AI offers further exciting potential to expand and improve renewable energy. For example, as a tool for scientific discovery, AI looks likely to accelerate the pace of innovation in key technologies such as photovoltaic (PV) solar modules, or battery storage. Improvements here could improve efficiency or performance, lower the technologies’ cost, or provide other tangible benefits. So, in essence, AI enables cleaner energy to power its own consumption. 

Why IDNOs are well-placed to lead

AI isn’t just for national infrastructure – it can drive efficiency at the distribution level, too. Independent distribution network operators (IDNOs) like Eclipse Power Networks are ideally placed to adopt AI tools quickly and with focus. With more agile structures than traditional distribution network operators (DNOs), IDNOs can trial and deploy targeted AI solutions across design, planning, operations and asset maintenance.

AI supports faster connections, the smarter adoption of existing networks, and predictive maintenance that minimises disruption and cost. At the planning stage, AI enables better demand forecasting and scenario modelling. And through data-led asset management, AI helps us move from reactive to preventative maintenance strategies.

However, while the power industry can already point to successes, it’s not alone in facing a critical, UK-wide shortage of AI skills. These add to the challenge of adapting to shifting and growing energy use, driven in part by AI’s increased demands. But there is increasing help, for example through the free tools, funding opportunities and knowledge-transfer partnerships supported by Innovate UK. By investing in skills, and by reframing AI as a strategic enabler across business functions, the power industry can continue to innovate in this area.

From condition to cure

Despite some alarmism about AI’s energy consumption, it has the potential to more than compensate for its own energy demand. As the IEA’s recent report finds, it could be instrumental in cutting costs, enhancing competitiveness and reducing emissions across the sector.

By using AI to optimise demand and generation, we don’t have to take a simple ‘add more to keep up’ approach. Intelligent AI-powered optimisation of smart grids and actively managed networks can do lots of the heavy lifting that the new hyperscale data centres will need, as well as powering a cleaner energy infrastructure that benefits everyone.