With AI ambitions putting fresh pressure on energy infrastructure, Taco Engelaar, Senior Vice President and Managing Director at Neara, argues that better grid visibility will be critical to turning bold data centre plans into reality.

Despite record-breaking funding and the policy commitments to match, AI ambitions in the UK and Europe face a reckoning. Concerns over ‘phantom investments’ have called into question whether bold plans will materialise on the ground – a concern amplified by OpenAI’s withdrawal of Stargate UK. As promising plans run up against the practical limitations of delivery, more projects may face delays or reassessment. A worrying ‘visibility gap’ is emerging between data centre design and operational reality.

However, it is not all bad news. While the ‘visibility gap’ exposes the real obstacles to data centre expansion and digital sovereignty in the UK and Europe, it also presents an opportunity: to better understand what is possible within the bounds of existing infrastructure, and to realign data centre strategy around practical delivery.

As rising geopolitical instability heightens the importance of digital sovereignty and control over compute, AI infrastructure cannot be considered separately from the physical systems that support it. By being honest about the limitations we face – and building with, rather than in spite of, them – the industry can begin to make more meaningful progress.

Reframing our approach to the grid

The first step is to stop treating existing infrastructure solely as an obstacle. There is no denying that ageing energy networks pose significant limitations for new data centre developments. Capacity constraints are real: new sites currently face waits of up to 10 years to connect to the grid in Europe. But this is not an insurmountable hurdle. The challenge is to navigate these limitations more intelligently and make better use of what is already available.

To do this requires accurate, asset-level insight, and a more physically realistic understanding of the grid. Right now, data centre planning often lacks both. Designs may account for static capacity constraints, when in reality there are fluctuations, flexibility, and untapped capacity within the grid. If navigated effectively, these could help unlock additional headroom to support new connections.

AI itself has a role to play in providing this insight. When used to support physics-based digital modelling, it can help map energy infrastructure, assess capacity, and simulate the impact of new connections before they are given the green light. These insights can be used to evaluate different options – such as adjusting capacity limits, harnessing different energy sources, or managing data centre flexibility – all of which could help speed up permitting and bring new sites online sooner.

Setting realistic expectations

Improving our understanding of existing infrastructure and optimising capacity is critical, but it is not a silver bullet. Some new transmission infrastructure will still be needed to make space for new data centres, no matter how effectively we manage the existing grid. While publicly unpopular, with current plans expected to cost up to £70 billion in the UK alone, a certain amount of ‘rewiring’ is essential if we are to deliver new infrastructure and support a more resilient, sustainable energy system.

This impact can still be managed. The same physics-based modelling that can help enhance our understanding of the grid can also be used to guide decision-making around network expansion. Understanding precisely where constraints lie, which assets are likely to age quickest, and where untapped capacity can be unlocked can help leaders make more informed and proportionate decisions about how much new transmission is needed, reducing unnecessary cost and delay.

Alongside new infrastructure, alternative power sources may also need to be considered. While renewables are rightly a priority for powering new data centres, our current networks are years away from being able to support a fully renewable system. Development will ultimately rely on a more diverse energy mix. Traditional energy players, such as Drax and Shell, are beginning to move into this space, creating an opportunity to use a broader range of energy solutions to support near-term data centre growth.

Meanwhile, apparent challenges elsewhere in the system could be turned to advantage. For example, channelling high levels of solar production into powering data centres could help mitigate the surplus expected during low-demand periods and reduce the risk of curtailment.

Breaking down the barriers to collaboration

Bringing data centre ambitions to life is a complex balancing act. To get it right, more coordinated planning and collaboration is needed. Data centre planning cannot continue to happen in silos. Knowledge and insights need to be pooled more effectively, so that developers, utilities, policymakers, and energy providers are working from a shared understanding of grid reality.

All stakeholders need clearer visibility of the grid. This means more effective data sharing and accessibility from day one, with insights fed into every stage of design. Rather than taking a reactive approach, effective grid modelling and assessment should be part of the conversation from the very start. Without this coordination, plans will continue to face avoidable friction.

As the gap between data centre planning and reality widens, the industry is at a critical point. AI infrastructure expansion is no longer simply a question of economic competitiveness; it is increasingly linked to digital sovereignty, energy resilience, and long-term security. Closing the ‘visibility gap’ is therefore becoming an urgent priority.

The industry needs to be realistic about the state of existing infrastructure, identify practical opportunities within current constraints, and work together to improve decision-making and design. Only then will ambitious data centre plans have a better chance of moving from concept to delivery.