Digital Event Horizon
Emerald AI, a startup based in Washington D.C., has developed an innovative solution to tackle the growing issue of energy consumption by data centers, leveraging NVIDIA's cutting-edge technology to create an AI factory flexibility platform that can dynamically adjust power usage during peak grid demand. The company's breakthrough, showcased in a recent trial conducted in Phoenix, Arizona, demonstrates the vast potential of this approach, which could revolutionize the way data centers operate and provide meaningful relief to struggling grids.
Emerald AI's innovative platform enables data centers to dynamically adjust their energy consumption during peak grid demand. The platform, developed with NVIDIA, Oracle Cloud Infrastructure, Databricks, and Salt River Project, has demonstrated a 25% power reduction over three hours while preserving compute service quality. Emerald Conductor acts as a smart mediator between the grid and data centers, orchestrating AI workloads to meet power grid demands. The approach helps cities avoid rolling blackouts, protect communities from rising utility rates, and make it easier for grids to integrate clean energy. The solution is crucial in preventing data centers from getting disconnected from the power supply during load shed events.
In an era where artificial intelligence (AI) is increasingly becoming a cornerstone of modern industries, from healthcare to finance, one pressing issue has come to the forefront – energy consumption by data centers. As the global demand for electricity continues to rise, concerns over grid capacity and reliability have become more pronounced. Amidst this backdrop, Emerald AI, a startup based in Washington D.C., has emerged as a pioneer in addressing this challenge through its innovative AI factory flexibility platform.
Developed in collaboration with NVIDIA, Oracle Cloud Infrastructure, Databricks, and the regional power utility Salt River Project (SRP), Emerald AI's solution leverages the vast capabilities of NVIDIA GPUs to enable data centers to dynamically adjust their energy consumption during peak grid demand. This breakthrough technology has been showcased in a recent trial conducted in Phoenix, Arizona, where the company demonstrated that its platform can reduce power consumption by 25% over three hours while preserving compute service quality.
At the heart of this innovation lies Emerald Conductor, an AI-powered system that acts as a smart mediator between the grid and a data center. This platform enables flexible energy usage by orchestrating various AI workloads across a network of data centers to meet power grid demands. Some jobs can be paused or slowed down, while others, like inference queries for an AI service used by thousands or even millions of people, cannot be rescheduled but can be redirected to another data center where the local power grid is less stressed.
This approach not only helps AI factories come online using existing power systems but also offers a means to modulate power usage, which could help cities avoid rolling blackouts, protect communities from rising utility rates, and make it easier for grids to integrate clean energy. According to Ayse Coskun, Emerald AI's chief scientist and a professor at Boston University, "Renewable energy, which is intermittent and variable, is easier to add to a grid if that grid has lots of shock absorbers that can shift with changes in power supply."
The significance of this breakthrough cannot be overstated. As the International Energy Agency projects that electricity demand from data centers globally could more than double by 2030, the need for innovative solutions like Emerald AI's is becoming increasingly pressing. The state of Texas has already passed a law requiring data centers to ramp down consumption or disconnect from the grid at utilities' requests during load shed events. In such situations, dynamic power reduction can be crucial in preventing data centers from getting disconnected from the power supply entirely.
Emerald AI is expanding its technology trials in Arizona and beyond, with plans to continue working with NVIDIA to test its technology on AI factories. The company's Phoenix demonstration, part of EPRI's DCFlex data center flexibility initiative, was executed collaboratively by Emerald AI, NVIDIA, Oracle Cloud Infrastructure (OCI), Databricks, and the regional power utility Salt River Project (SRP). This landmark trial validated the vast potential of an essential element in data center flexibility, as highlighted by Anuja Ratnayake, who leads EPRI's DCFlex Consortium.
In a recent trial conducted on May 3, 2025, with high air-conditioning demand in Phoenix, SRP's system experienced peak demand at 6 p.m. During the test, the data center cluster reduced consumption gradually over a 15-minute ramp down and maintained the 25% power reduction for three hours before ramping back up without exceeding its original baseline consumption. This remarkable demonstration showcased Emerald AI's ability to dynamically adjust energy usage while preserving compute service quality.
The company's innovative approach is also supported by the Emerald Simulator, which accurately models system behavior to optimize trade-offs between energy usage and AI performance. Historical grid demand from data provider Amperon confirmed that the AI cluster performed correctly during the grid's peak period. Furthermore, a comparison of Emerald Simulator prediction of AI GPU cluster power with real-world measured power consumption demonstrated the accuracy and effectiveness of this tool.
Looking ahead, Emerald AI is poised to forge an energy-resilient future by harnessing NVIDIA technology to create more flexible, responsive, and sustainable data centers. As Varun Sivaram, founder and CEO of Emerald AI, aptly put it, "We can make data centers controllable while assuring acceptable AI performance. AI factories can flex when the grid is tight — and sprint when users need them to."
In conclusion, Emerald AI's innovative solution has demonstrated that harnessing NVIDIA technology can unlock new possibilities for data center flexibility and sustainability. As the world continues to grapple with the challenges of energy consumption by data centers, this breakthrough offers a beacon of hope – a chance to reimagine data centers as helpful resources in managing grid stress and promoting clean energy integration.
Related Information:
https://www.digitaleventhorizon.com/articles/Unlocking-the-Full-Potential-of-AI-Factories-How-Emerald-AI-is-Harnessing-NVIDIA-Technology-to-Tackle-Grid-Stress-deh.shtml
https://blogs.nvidia.com/blog/ai-factories-flexible-power-use/
https://firmus.co/company-news/firmus-and-eaton-showcase-how-ai-factories-can-support-the-grid
Published: Tue Jul 1 09:41:19 2025 by llama3.2 3B Q4_K_M
