Artificial intelligence (AI) continues to empower modern society’s daily habits and is integrated throughout our lives, with accessibility at our fingertips. Progress has escalated at such a rapid scale that connectivity is no longer a luxury; it is essential.
Behind the technology is the need for the energy infrastructure to support it. The countries with the largest concentration of data centres, including China and the United States, are growing at scale, largely supported by renewable energy generation and the ability to connect to the grid. Innovative computer chips have started to decrease the amount of energy used by AI in data centres, but this hasn’t reached its full potential. In the UK and some European markets, grid connection backlogs are limiting AI developments. For example, the Global Infrastructure Investor Association has reported that grid reinforcement in the UK can take up to 15 years, a statistic considerably slower than the rate of AI advancement, adding to an ever-increasing bottleneck.
What’s the challenge?
Today, searching on an AI platform places an increased demand on the electricity grid. Evidence shows that one search on AI consumes 10-12x the electricity compared to a traditional Google search. Of course, one search is not representative of how the average person uses a large language model with a number of prompts being used, this translates to 2.9 kilowatt-hours of electricity – enough to cycle over 100 miles on an e-bike or drive around 10 miles in the average electric vehicle.
A recent publication from the US Department of Energy has warned that a crisis is looming due to unprecedented growth in data centres fueled by AI consumer demand. This was mirrored in the UK by a policy paper from the Department of Energy Security and Net Zero that states “Significant reinforcement and build out of the distribution network will also be required to support the electrification of sectors projected for the decades ahead, as well as to accommodate new demand in some locations for growing infrastructure and industrial uses, such as data centres and transport hubs”. This is a global issue with energy grids stretched far beyond capacity. The International Energy Agency has reported that electricity demand from data centres is set to double by 2030 to around 945 terawatt-hours (TWh) – the equivalent of the yearly electricity consumption of Japan. AI technology is the biggest driver of this vast increase.
Criticality of the infrastructure
Energy infrastructure plays a crucial role in supporting the power used by AI platforms, and globally, the infrastructure requires major investment to cope with the growing demand for AI. To ensure there is a reliable power flow, transformers sit at the epicentre of data centres’ operations. They step down high-voltage electricity from the grid, ensuring a stable power supply to critical AI hardware such as graphic processing units (GPUs). By adopting a holistic transformer monitoring strategy, data centres can gain the visibility and foresight needed to prevent failures, extend asset lifespan and maximise uptime.
Data centres are now increasingly being classified as Critical National Infrastructure and contain a huge level of redundancy cluding, UPS systems, backup generators and spare transformers. However, due to the huge reliance on data centres, the role of monitoring the transformers and their backup systems is essential. This was highlighted in a UK Government case study, which states, “The vulnerability of this infrastructure was starkly illustrated in July 2022, when two separate data centres serving an NHS trust failed during a heatwave. The incident took down most clinical IT systems at three hospitals, as well as related community services. The disruption to patient care was massive and widespread, forcing the Trust to spend £1.4 million in unplanned technology costs to respond.”
Data centres need to invest in monitoring capability for critical assets to enable a greater understanding of when their most critical assets require maintenance to avoid unplanned outages and extend transformer lifespan.
Industry challenges
For many network and data centre operators, the major challenge is addressing the widespread strategic risk. Transformer failures are no longer isolated incidents; they are now becoming a major threat that could destabilise the electricity grid.
The added pressures on the grid have major risks, including:
- Rising risk of blackouts:
The U.S. Department of Energy warns of up to 800 hours of outages per year by 2030, which equates to a full month of blackouts, if our energy infrastructure fails to adapt.
- Transformer supply bottlenecks:
Transformers have been a major supply chain challenge, and lead times for new transformers globally are stretching up to seven years.
- Unprecedented demand from AI and electrification:
Modern technology is consuming significant levels of energy compared to traditional sources. However, in certain regions, including the UK and Ireland, there is a bottleneck in connecting to the grid, slowing down innovation.
- Subsequent consequences of failure:
A single large-scale transformer failure can create huge consequences. For example, the Heathrow Airport incident in 2025 caused by a transformer failure had a significant economic impact. The estimated daily financial loss reached £20 million, and the loss of tourism cost £4.8 million per day as flights were cancelled for more than two days.
How Camlin Energy is empowering tomorrow’s grid
For data centres to be commercially viable, they must be operational 24/7 and available at peak efficiency levels to cope with the power generated from AI technology. Avoiding downtime is the highest priority, and a reliable electricity supply is critical. A transformer is the link between data centres and the electricity grid for power generation and distribution.
Camlin Energy is trusted by leading data centres globally to strengthen the energy infrastructure and protect their most critical assets with end-to-end transformer monitoring solutions, allowing asset managers to gain deeper insights into the health, performance and risks that their transformers face, supporting data centres to achieve maximum uptime.
Importance of implementing a transformer monitoring strategy
Implementing a holistic transformer monitoring strategy is essential to safeguard against risks, including:
- Preventing catastrophic failures:
Real-time monitoring can detect early signs of issues (e.g., insulation degradation and overheating), enabling proactive maintenance to prevent costly disruptions and downtime.
- Mitigate supply chain risks:
Monitoring systems help extend transformer lifespan through timely interventions, reducing reliance on scarce replacements.
- Enhance grid resilience:
Advanced monitoring solutions provide data-driven insights into transformer health, enabling operators to prioritise upgrades, optimise maintenance schedules, and strengthen grid reliability against demand.
- Regulatory and stakeholder pressure:
Regulators and customers demand greater reliability and sustainability. Having a proactive monitoring strategy is crucial to support the modernisation of the grid and to meet environmental, social, and governance goals.
In conclusion
AI is continuing to grow at a rapid pace, pushing data centres to operate at peak capacity levels and beyond. This requires a robust infrastructure system to support innovation, ensure reliability and resilience. We can see improvements are being made with more efficient GPUs to save energy in the future, but today, there is a major challenge, as one AI search consumes over 10x as much electricity.
Reports indicate that the energy used by data centres is set to double by 2030. Operators must adopt a holistic transformer monitoring solution to address the threats posed by AI that is impacting critical infrastructure, supply chain bottlenecks, and the increasing demand on data centres.
Asset monitoring offers a solution that enhances grid reliability, operational performance and reduces risks. Investing in these solutions now is a strategic move to ensure the energy network is future-proofed to cope with the demand from AI and growing pressures from data centres.
