The UK's AI ambitions is there but the sector needs more talent

The Data Centre Alliance's (DCA) new positioning paper, Enabling AI at Scale in the UK, makes a case that should resonate beyond those who make policy: the UK's AI ambitions are not really a software problem. They're a physical infrastructure problem: grid capacity, planning consent, cooling engineering and substations. The DCA argues that without coordinated policy across DSIT, DESNZ and MHCLG, the UK risks having world-class AI ambition with nowhere fast enough to run it.

What the paper doesn't highlight, but what anyone recruiting into this sector already knows, is that the same bottleneck applies to people. You cannot build or run data center without the right talent, and the UK like almost all countries is lacking in this area.

The numbers back this up. Data centres are now the second most in-demand construction sector in the UK, and Turner & Townsend has found that construction cost inflation is expected to climb through 2026 and 2027 as data centre and AI-related infrastructure demand squeezes an already stretched market, with over 70% of markets reporting tightening contractor capacity and 87% reporting shortages specifically in the mechanical, electrical and plumbing trades that data centres depend on. Turner & Townsend's data also situates this against a UK labour market where 75% of UK construction markets are already experiencing shortages.

It isn't just construction. Salute's State of the Industry 2026 report found shortages of specialist operations talent, limited training pathways and stiff competition from hyperscalers among the biggest barriers to AI readiness, with 74% of leaders saying they aren't fully prepared to support AI workloads today. As one industry partner put it, "skills will be an ever-growing pressure on the UK data centre industry through 2026, and the sector can't keep pace without widening the talent pool."

This is where the DCA's recommendations and the recruitment industry's day job converge. If AI Growth Zones are to function as genuine "infrastructure delivery ecosystems" rather than branding exercises, and if front-loaded development is to work as the DCA proposes, then workforce planning needs to sit alongside power feasibility studies and planning consent from day one - not as an afterthought once a site is energised.

For specialist recruiters, the message is straightforward: the DCA is describing a multi-year, nationwide build-out, construction trades, MEP engineers, critical facilities operations, and increasingly the AI-specific skills needed to run inference at the edge. Firms that can supply talent pipelines aligned to that timeline, rather than reacting to it, will be the ones data centre developers turn to first.