At Data Centre London, the conversation has moved on from whether AI will drive growth – the industry is now firmly focused on how to deliver at pace, while navigating the practical realities of power, sustainability and regulation. Across the sessions, one theme came through clearly: innovation is accelerating, but the constraints haven’t changed.
Sovereignty becomes real
AI sovereignty is quickly shifting from a strategic idea to a delivery requirement.
With EU regulation emerging, 
operators – and critically, their supply chains – will need to demonstrate compliance across multiple aspects of resilience and control. This signals a step change. Sovereignty will no longer sit at the policy level; it will need to be embedded into how data centres are designed, built and operated from the outset.
Sustainability is a balancing act
Sustainability discussions moved beyond ambition into trade-offs, particularly the tension between water and energy. Reducing water usage can increase power demand, forcing operators to actively balance PUE against WUE. There is no universal solution – each site is shaped by climate, water availability and infrastructure. What is clear is that early decisions matter more than ever. From initial due diligence through to long-term maintenance strategies, small choices can have lasting impacts on efficiency.
Evolving models, familiar blockers
The market continues to explore new delivery models – from GPU-as-a-service to more distributed, “neo cloud” platforms. But despite this innovation, the same challenges dominate: power, permits and talent.
Even in regions with relative advantages, such as France’s nuclear capacity, delivery timelines are still heavily influenced by planning and regulatory constraints. The expectation is growth – but with a degree of consolidation as the market matures.
AI’s dual role in energy
AI is not just a driver of demand – it’s starting to shape how energy is managed.
While the immediate challenge is meeting the scale of additional consumption, there is growing focus on using AI to optimise performance. In the short term, this means improving efficiency within buildings; longer term, it could extend to smarter, more responsive energy sourcing at a systems level.
This dual role – both consumer and optimiser – is becoming a defining dynamic.
Designing for density
Rising rack densities are reshaping data centre design. As compute demand increases, the proportion of space dedicated to power and cooling infrastructure is growing, reducing the relative footprint of white space.
In response, operators are leaning into more flexible, modular solutions – enabling different configurations as technologies evolve. The ability to adapt is becoming just as important as the ability to scale.
The push for transparency
Sustainability remains high on the agenda, but the industry still lacks consistency in how it’s measured.
Without standardised KPIs, it’s difficult to demonstrate performance or build trust – particularly with local communities who are increasingly influencing planning decisions. The challenge now is not just to improve outcomes, but to prove them clearly and consistently.
Navigating the path to scale
Across all discussions – from hyperscale strategy to infrastructure design, the same set of pressures continues to shape decisions: sovereignty, sustainability, power availability, permitting and access to talent. They don’t operate in isolation, and they won’t be solved quickly. The takeaway is a pragmatic one: AI may be driving demand at unprecedented speed, but delivering the infrastructure behind it still depends on coordination, capability and careful planning.
About Soben and Accenture
Soben is part of Accenture’s Infrastructure and Capital Projects practice, bringing together hands-on data centre delivery expertise with Accenture’s digital and industry capabilities across the project lifecycle.
For more insight into how the sector is evolving, explore our latest Data Centre Trends Report, or learn more about Accenture’s Data Centre Services.