Nvidia-backed data center orchestration platform Emerald AI has announced a pilot program with California municipal utility Silicon Valley Power (SVP) in Santa Clara, California.

The pilot project will see Emerald AI work with several SVP customers to demonstrate how its Emerald Conductor platform can help data centers support grid flexibility. According to the companies, the pilot will evaluate how flexible operations can help safeguard reliability, make more efficient use of existing infrastructure, and inform planning for bringing forward phased energization as data center load ramps on the SVP system.

The company’s flagship product, the Emerald Conductor program, serves as a mediator between the grid and data centers, orchestrating AI workloads in real time. This enables data centers to dynamically adjust their energy consumption, supporting grid stability while ensuring acceptable AI compute performance.

The company will test the technology at a multi-megawatt facility at a Centersquare data center in Santa Clara. The company operates five data centers in the city.

“Santa Clara is at the center of the next wave of AI and digital infrastructure investment, and Silicon Valley Power is committed to supporting that growth responsibly,” said Nico Procos, electric utility director of Silicon Valley Power. “This pilot will help us evaluate practical tools to protect reliability and affordability for our customers while supporting flexible, efficient planning for future load growth. I want to acknowledge our partners, Emerald AI and Nvidia, and also the staff at SVP for their innovative thinking, led by our chief operating officer, Chris Karwick.”

“Data centers can be more than static electric loads – they can become active partners to the grid,” added Varun Sivaram, founder and CEO of Emerald AI.

The pilot is Emerald AI’s fourth, having completed two in the US, in Phoenix and Chicago, and one in the UK in partnership with the National Grid. According to the company, all three demonstrations have proved the efficacy of its technology, with its latest in the UK successfully simulating more than 200 real-time “grid events” to test the Conductor’s ability to dynamically adjust the data center’s power consumption. According to the company, the platform successfully adjusted power usage to the requested level, cutting power demand by up to 40 percent without impacting critical workloads.

The trial was expected to serve as an operational blueprint for its first commercial project in Virginia. The company announced plans for the project in October, with the orchestration software set to be deployed at Nvidia’s under-construction 96MW Aurora data center in Manassas, Virginia.

The company is a member of the Nvidia Inception program for startups and is also backed by Google’s chief sustainability officer, Kate Brandt. The firm has raised more than $67.5 million to date, over three funding rounds. Its latest, a seed extension, raised $25m earlier this year.