Phil Cullerton, VP Services EMEA at Vertiv
Artificial intelligence (AI) is often described in terms of algorithms, computing power, and data. Yet the pace of AI adoption is now shaped as much by the physical constraints of critical digital infrastructure as by software innovation.
Across Europe and beyond, surging demand for AI-ready capacity is spurring rapid advancements in power networks, cooling systems, and engineering expertise to support it. Operational teams are quickly adapting to bridge the gap between technical possibilities and deployment realities.
This tension sits at the centre of modern digital economies. Nations want to attract AI investment, scale sovereign capabilities, and create a competitive environment for research and innovation. But AI workloads place demands on physical systems that were not designed for them. Understanding how critical digital infrastructure is built, maintained, and governed is becoming a core part of digital policy.
AI is no longer just a compute challenge
The growth in model size and training cycles has led to a step-change in power density. Many organisations now operate workloads that draw tens of kilowatts per rack, with some designs exceeding that threshold. These environments behave differently from traditional cloud or enterprise deployments. They generate sharper thermal fluctuations, more sustained pressure on cooling and a higher risk of system drift when supporting equipment is not calibrated accurately.
This means AI readiness is not determined only by how many graphics processing units (GPUs) or accelerators an operator can purchase. The limiting factors are increasingly practical: grid access, commissioning quality, thermal stability, fluid chemistry and the availability of multidisciplinary engineering talent.
The economic consequence is clear. Regions that cannot deliver the operational backbone for high-density environments will find themselves at a disadvantage, regardless of policy ambition or investment incentives.
Regulation is reshaping operational expectations
Governments are introducing new requirements around energy performance, transparency and environmental responsibility. The European Union’s revised Energy Efficiency Directive mandates reporting on energy and water performance for qualifying data centres. Germany’s Energy Efficiency Act requires specific power usage effectiveness (PUE) thresholds and waste heat reuse obligations for new facility builds.
These rules reshape how sites must be designed, monitored and maintained. They also increase the operational burden on teams responsible for compliance. Meeting legal requirements is no longer a matter of documentation. It depends on real-time data, continuous performance validation and predictable, well-managed operations.
Digital economies that want to remain competitive need to recognise that AI growth and regulatory complexity now rise in parallel. Strong operational frameworks will be needed to reconcile the two.
Services and skills now have economic weight
A data centre designed for AI density requires engineers who understand mechanical cooling, electrical power, fluid loops, thermal behaviour and digital monitoring. Few markets can supply this blend of expertise at scale. CRG Tec Recruitment reports that more than half of operators already struggle to fill critical roles, and the demand curve continues to rise.
This shortage has direct economic consequences. Without skilled operators, even well-built facilities struggle to achieve planned availability and efficiency targets. Projects slow down, delays compound and AI investment shifts toward regions with more reliable operational capability.
Digital economies often discuss skills gaps in the context of software development or data science. Increasingly, the skills shortage underpinning physical infrastructure is just as critical to competitiveness. AI strategy must therefore include investment in engineering capability, training pathways and cross-skilling programmes.
Operational resilience is becoming national infrastructure
Digital policy for the last decade focused on cloud adoption, broadband, cyber security and data governance. As AI workloads reshape energy use and environmental impact, operational resilience now matters as much as compute capacity.
The infrastructure supporting AI does not exist in isolation. It affects national power networks, local planning processes, waste heat reuse schemes, water consumption and industrial policy.
A region that cannot provide predictable operational conditions for high-density environments risks limiting growth in research, financial services, advanced manufacturing and public-sector digitalisation.
Services, once seen as a technical afterthought, now sit at the centre of competitiveness. They determine whether facilities can scale responsibly, whether regulatory standards are met and whether investors can rely on infrastructure performance.
ALSO READ:IHS REPORTS $1.58BN REVENUE AS GROWTH HITS 10.1%
A shift from reactive to lifecycle models
A noticeable change across modern facilities is the move from reactive problem-solving to integrated lifecycle strategies. Instead of treating installation, commissioning, maintenance and optimisation as separate activities, leading operators link these stages into a continuous process.
There are reasons for this shift. Small deviations in vibration, compressor behaviour or fluid balance often indicate deeper system issues. Detecting and addressing these signals early can prevent costly failures.
Lifecycle strategies also help integrate regulatory requirements into day-to-day practice. Heat reuse obligations, PUE thresholds and environmental reporting all benefit from long-term operational planning rather than piecemeal compliance efforts.
Digital economies that understand and adopt lifecycle models are better equipped to integrate AI efficiently and predictably.
Real-world lessons from AI expansion
Several deployment patterns have emerged across Europe and beyond.
In remote northern regions, operators preparing for AI-ready facilities discovered that commissioning and spare-parts logistics had to be localised. Without on-the-ground support, timelines became unpredictable and operational risk increased.
In Germany, new energy efficiency rules forced operators to redesign thermal strategies, integrate waste heat reuse early and track energy performance more rigorously. This alignment between regulation and operation has already influenced how new sites are planned.
In Southern and Eastern Europe, organisations transitioning to high density have used monitoring tools and cross-skilled teams to stabilise operating costs and reduce incident rates.
These examples show that infrastructure success increasingly depends on operational maturity, not just capital investment.
What digital economies must prepare for next
The physical demands of AI are likely to increase as adoption and scale expand. Liquid cooling and hybrid approaches are gaining traction across both hyperscale and enterprise environments. Waste heat reuse is moving toward regulatory expectations in several European markets. Power distribution upgrades are influencing national energy and industrial policy.
Regions that want to attract AI investment will need reliable access to engineering talent, clear policy frameworks, predictable grid infrastructure, and a mature approach to lifecycle operations. The geopolitical race for compute capacity will be influenced as much by operational resilience as by capital availability.
The global conversation around AI often focuses on models, safety and ethics. These topics matter. But there is another layer shaping the future of AI adoption – the infrastructure services responsible for keeping physical systems stable. The countries and regions that understand this operational reality will be the ones best prepared for the next phase of digital transformation.
