UK-first trial of AI Grid Technology Successfully Demonstrates the Ability of Data Centres to Adjust Power Needs

Emerald AI, EPRI, National Grid, Nebius, and NVIDIA today announced a landmark, UK-first demonstration proving high-performance AI infrastructure can operate as a flexible, grid-responsive asset. Using Emerald AI’s software, Emerald Conductor, to manage a cluster of NVIDIA GPUs, the trial validated that artificial intelligence data centres can dynamically adjust power consumption in response to real-time signals, without disrupting critical workloads.

Today, most large data centres are treated as fixed, “always on” demand. As AI use grows and more facilities seek to connect, the idea that data centres are a fixed demand could increase network constraints and lengthen connection times. Demonstrating the ability of data centres to flex their power demand shows how they can ease constraints and unlock grid capacity, rather than be an extra source of inflexible demand.

Live trial: from half time teas and system management

The trial tested Emerald AI’s software on a NVIDIA Blackwell Ultra cluster, a bank of 96 high performance GPUs at Nebius’s new data centre in London.

Over five days in December 2025, more than 200 real-time simulated “grid events” were sent to the site to test Emerald’s Conductor’s ability to dynamically adjust the data centre’s power consumption.

Each time, Emerald AI’s platform successfully adjusted power use to the requested level, cutting demand by up to 40% while critical workloads continued to run as normal.

Key results included:

• Peak smoothing: Successfully reacted to spikes in demand during half time of major football matches, proving AI can counterbalance sudden consumer electricity surges.
• Long-duration shifting: Followed load-reduction requests for up to 10 hours, helping the grid navigate periods of low wind or extreme heat.
• System management: Simulated a system stress event, shedding 30 percent of its load in roughly 30 seconds to help maintain grid resilience.

As the UK prepares for over 6 GW of data centre deployments on the grid by 2030, this technology could enable AI data centres across the UK to add more than 2 GW of capacity back to the grid when needed.

The success of the trial demonstrates that AI data centres can move from being a source of electricity constraint to a controllable grid asset. By flexing demand in real time, they can help manage peaks, make better use of existing infrastructure, and support the connection of different sources of energy to the grid.

Steve Smith, President, National Grid Partners said:

"As the UK’s digital economy accelerates, there’s concern that data centres could add pressure to an already constrained system. This trial proves the opposite can be true. High‑performance data centres don’t have to place additional strain on the grid. With our partners, we’ve shown they can be connected and managed without major new network capacity, flexing their power up or down in real time to support the whole system. This approach will enable us to connect significant new demand more quickly and, help to lower network charges for customers over time."

Dr Varun Sivaram, Founder and CEO of Emerald AI, said:

"This trial demonstrates that AI infrastructure can be a dynamic force for the grid. With dozens of realistic AI workloads running simultaneously, we delivered fast emergency curtailment and sustained, precise peak reduction. The same approach we validated here can be applied to much larger AI factories, as the industry scales."

David Porter, EPRI Vice President of Electrification and Sustainable Energy Strategy said:   

"With each DCFlex demonstration, we learn more about operationalising data centre flexibility to support AI-driven load growth while ensuring the reliability of the electric system. Field demonstrations are a unique element of DCFlex and are designed to accelerate the adoption of solutions through real world validation, showing how AI infrastructure can become a flexible, grid supporting resource."

Daria Mukhortova, Head of Sustainability at Nebius, said:

"For Nebius, optimizing the energy footprint of AI workloads is a core priority in how we build our AI cloud, from physical infrastructure to software. As AI grows, its interaction with the power system is increasingly becoming a focus for the industry. Trials of new technologies such as this indicate how AI might operate more flexibly as part of the grid."

Josh Parker, Head of Sustainability at NVIDIA said:

"The era of the 'always-on' AI factory is evolving. This trial proves that NVIDIA-powered infrastructure can act as a grid-aware asset, modulating demand in real-time to support stability. By making AI workloads responsive, we accelerate deployment while reducing the need for costly grid upgrades."

The performance data will now be shared with industry, regulators and policymakers to influence future rules for “power flexible” data centre connections, including how such facilities could qualify for faster and larger connections where they agree to flex on request.

The London trial serves as the operational blueprint for Aurora, a nearly 100MW power-flexible AI Factory, operated by NVIDIA in Virginia. By establishing measurement-based evidence for AI flexibility, the partners aim to set new global standards for interconnection.

Emerald AI, a National Grid Partners and NVIDIA NVentures portfolio company, is integrating these capabilities into the NVIDIA Omniverse DSX Blueprint for gigascale AI Factories. This "grid-aware by design" approach ensures that as the UK races toward AI leadership, the energy infrastructure remains resilient, affordable, and sustainable.