National Grid and Emerald AI test flexible data centre power model
Subscribe to our free newsletter today to keep up to date with the latest construction and civil engineering news.
National Grid and Emerald AI are launching a real-time trial to prove that artificial intelligence-powered data centres can adjust their energy usage dynamically to relieve pressure on the UK’s power grid. The pilot, scheduled for late 2025, aims to show that smarter infrastructure could help reduce the need for major new civil engineering works, opening a path to faster deployment of digital infrastructure.
The demonstration will use Emerald AI’s platform, Emerald Conductor, to manage energy loads at a UK-based data centre equipped with advanced NVIDIA GPUs. If successful, the technology will show how AI-driven facilities can act as responsive energy consumers, reducing demand during peak periods. By doing so, data centres could become grid assets rather than burdens, helping to unlock latent capacity in the existing electricity network.
Flexible energy use reshapes traditional infrastructure timelines
For civil engineers and construction firms, the impact could be significant. Traditionally, connecting a large-scale data centre to the power grid triggers a lengthy process of grid assessments, civil works and equipment installation, often involving months or years of planning. These projects typically require the construction of substations, high-voltage cabling, and reinforced foundations, along with extensive cooling systems to manage heat generated by server operations.
Flexible demand changes that equation. By allowing facilities to temporarily reduce their energy consumption, the need for large-scale reinforcements to local grid infrastructure can be avoided or delayed. This could reduce or reshape the scope of traditional civil engineering works associated with grid upgrades, while enabling faster approvals for energy-hungry projects.
Engineering implications for new data centre builds
Engineering consultancies and construction contractors involved in grid-tied infrastructure are watching closely. Building flexibility into data centre architecture requires new design principles. Electrical systems must accommodate dynamic loads, with greater reliance on automated monitoring, distributed energy resources, and real-time control systems.
Cooling infrastructure must also be designed with demand variability in mind. This may include adaptive cooling loops, modular thermal storage, or even integration with district energy systems. On the ground, these adaptations affect how civil engineers lay out cable routing, trenching, building foundations and access ways for utility vehicles.
The design and construction of substation pads, cable ducts and support pylons will also need to be modified. While total project volumes may decrease in some regions, design complexity is likely to rise. Contractors may face pressure to deliver more agile infrastructure that meets the needs of a system that can scale up or down in real time.
Impacts across planning, skills and project delivery
The shift toward smart energy management has ripple effects across the construction ecosystem. Planning frameworks will need to evolve to consider energy flexibility as a factor in project approvals. Planning authorities may be more willing to approve data centre proposals if developers can demonstrate that facilities will not contribute to peak grid congestion.
From a workforce perspective, there will be rising demand for civil engineers skilled in electrical integration, smart infrastructure, and cooling system design. Recruitment challenges are likely to emerge, especially in regions already experiencing a tight labour market for infrastructure projects. Materials sourcing could also shift as developers opt for prefabricated components that support modularity and quick deployment.
Flexible power strategies will also influence how projects are sequenced. Civil works may need to follow more agile methodologies, with phased delivery schedules that allow for faster grid interconnection without waiting for full build-out. This will require closer coordination between developers, grid operators and contractors to ensure that core infrastructure aligns with evolving energy profiles.
A quiet evolution with potential for wide influence
While the Emerald AI trial may not grab headlines outside the energy and construction sectors, the underlying principle has broad application. If successful, it could signal a transition in how nations approach the civil infrastructure that underpins digital growth. For developers, engineers and investors, the message is clear. Infrastructure that flexes may be just as valuable as infrastructure that scales.
Sources: