AI & Technology

The UK’s AI ambitions need a smarter energy strategy

By Chris Bowden, founder and CEO of SQE (formerly Squeaky Energy) – an energy supply platform exclusively for industrial and commercial (I&C) users

The UK’s ambitions for artificial intelligence (AI) and digital infrastructure are rapidly bringing a new energy challenge into focus. As the government pushes for Britain to become an AI superpower, electricity demand from data centres is expected to rise significantly, with estimates suggesting around 6GW of additional capacity could be required by 2030. 

Much of the recent debate has centred on whether meeting the rapid growth in electricity demand from AI and data centres will inevitably require more gas-fired generation. But treating more gas as the only viable route to digital growth risks missing a much bigger opportunity. 

The UK will undoubtedly need more firm power if it wants data centres and AI to grow at the pace the government is talking about. Data centres are large, concentrated and reliability-sensitive loads, and they need electricity every second of every day, regardless of whether the wind is blowing or the sun is shining.

However, that does not automatically mean the answer is simply to build large amounts of new gas generation. 

Gas is likely to remain part of the electricity system for some time, particularly at the margins when renewable output is low. But private gas-fired power stations built specifically for data centres should be viewed as an absolute last resort, not the default model. Otherwise, we risk creating a parallel power system where well-funded digital infrastructure secures dedicated firm capacity, while the wider electricity network – and ultimately other consumers – carry the cost and complexity of balancing the grid. 

The real – and largely overlooked – opportunity lies in reimagining data centres, not as passive electricity consumers, but as active participants in the UK’s energy transition. 

From passive loads to active energy assets 

The conventional view is that data centres require perfectly flat electricity demand around the clock. Certainly, some critical computing functions do. But not all computing workloads are equal. 

AI training, batch processing, data backups and certain other computational tasks can often be shifted in time – or even location – without affecting the end user. That flexibility creates an opportunity that remains largely untapped. 

Rather than treating data centres simply as large industrial loads, operators should increasingly be designing them as active energy assets.  

That starts with intelligent siting, locating facilities where the electricity network has available capacity. It also means securing long-term clean power through power purchase agreements, integrating battery energy storage systems, incorporating onsite generation like solar and wind where possible, and, crucially, responding dynamically to conditions on the electricity system.  

If operators can increase demand when renewable generation is abundant and reduce non-critical computing during periods of system stress, they can materially reduce reliance on fossil-fuelled generation without compromising resilience. 

Delivering this kind of flexibility will also require much greater transparency. Operators increasingly need visibility into the energy consumption and carbon impact of individual AI workloads so they can make informed decisions about when and where to run non-critical computing. Without that level of operational insight, demand-side flexibility risks remaining more of an aspiration than a practical tool for supporting the electricity system. 

As the UK’s electricity mix continues to shift towards renewables, this flexibility becomes increasingly valuable. The traditional model of industrial energy supply – built around purchasing stable baseload power through relatively liquid forward markets – is gradually disappearing. As dispatchable generation declines and more electricity is sold through mechanisms such as Contracts for Difference (CfDs), forward market liquidity is becoming more constrained, creating a more volatile and complex environment for large energy users. 

The future lies in a more integrated system that combines grid connectivity, renewable generation, battery storage, flexible demand and intelligent control into a coordinated whole. 

Battery storage will play a central role in managing short-term volatility and enabling greater penetration of renewable generation. But storage alone is not enough, and demand-side flexibility will become equally important.  

Few electricity consumers are better placed to provide that flexibility than modern data centres. 

Looking beyond “100% renewable” 

The industry has become accustomed to annual “100% renewable” claims, often backed by monthly or annual renewable energy certificates like REGOs. While these demonstrate investment in clean generation, they do not necessarily reflect what is happening physically on the electricity system. 

A data centre can make a claim to run on 100% renewable electricity while still increasing demand at exactly the times when there is no clean energy on the system, and the grid is relying on gas-fired generation. The issue is not simply what energy is bought over a year, but when electricity is actually consumed.  

The real test should therefore become contractual linkage: What electricity is being consumed? When is it being consumed? And what does the data centre operator do when the electricity system is under stress? 

These questions matter far more than annual accounting exercises. Some operators are already beginning to explore more sophisticated approaches, including time-based matching of renewable generation with electricity consumption, and dynamically shifting workloads in response to grid conditions. Combined with battery storage, advanced control systems and co-located renewable generation, these approaches move the concept of a “green data centre” beyond marketing language towards genuine system benefit. 

Avoiding a parallel power system 

There will inevitably be situations where onsite gas generation is justified, particularly where resilience requirements are exceptional, waste heat can be recovered efficiently, or local grid capacity is genuinely unavailable. 

But encouraging widespread behind-the-meter gas-fired generation as the principal route to expanding AI infrastructure would represent a significant policy failure. If the only practical way to build data centres is to attach gas-fired power stations to every campus, then we are effectively acknowledging that the national electricity system cannot support one of the country’s priority industries. 

A more effective policy would attach conditions to priority planning or accelerated grid connections. Operators should be expected to demonstrate what firm clean power they have secured, what battery storage and flexibility capabilities they bring, how they intend to reduce demand during system stress events, and what steps they are taking to avoid simply adding to peak demand. 

The objective should not only be to provide more electricity, but to ensure that major new electricity consumers strengthen the resilience and flexibility of the wider system.  

A different energy future 

Data centres undoubtedly create challenges for the UK’s electricity system. Demand could increase several-fold over coming decades, placing additional pressure on networks and generation capacity if managed poorly. 

But this growth does not have to undermine the UK’s net zero ambitions. The more important question is not whether data centres pose a challenge – they clearly do at face value – but how we choose to integrate them into the wider energy system. 

If they continue to operate as passive demand, they will increase system stress and deepen reliance on gas generation. But if operators, policymakers and electricity markets evolve together, data centres could become one of the key enabling technologies of the clean energy transition itself. 

Ultimately, the real policy question is not whether the UK needs more gas, it’s whether data centres should be expected to contribute to system flexibility and resilience.  

If the answer is yes, then the UK’s fastest-growing electricity consumers could become something far more valuable than simply another source of demand. They could become partners in delivering a cleaner, smarter and more resilient energy system. 

Related Articles

Back to top button