Emerald AI announced its official launch, presenting the results of its first commercial demonstration and $24.5 million in seed funding led by Radical Ventures, with participation from NVentures (NVIDIA’s venture capital arm), AMPLO, CRV, and Neotribe. Emerald AI also receives support from individual investors, including Google Chief Scientist Jeff Dean, former Secretary of State John Kerry, former Australian Prime Minister Malcolm Turnbull, AI expert Professor Fei-Fei Li, and Kleiner Perkins Chairman John Doerr.
The company’s software aims to address the challenge of the electric power system meeting AI’s increasing energy demand. New US data centers are expected to need 50-100 GW of additional power by 2030, but the lack of available power grid interconnections may hinder AI innovation and U.S. competitiveness. And the North American grid reliability regulator has indicated that this increased demand could affect power system stability, while substantial investments to develop electricity infrastructure over the next decades may increase energy prices for American communities.
Emerald AI gathered a team headed by Dr. Sivaram, who is a former Fortune 500 C-suite energy executive, senior U.S. diplomat, Rhodes Scholar, and physicist. And other team members include Boston University Professor Ayse Coskun, an expert in flexible AI and high-performance computing, Shayan Sengupta, who has extensive hyperscale cloud and AI leadership experience from industry giants like Amazon and Intel, and Aroon Vijaykar, a former energy executive at Sunrun and chief executive of AEE Solar.
The Emerald AI Conductor software platform manages AI workloads in real-time, enabling data centers to dynamically adjust their energy consumption to support grid stability while maintaining acceptable AI compute performance. This software-based management of energy consumption enables AI data centers to avoid lengthy grid interconnection queues.
Emerald AI’s technology has shown effectiveness in significant real-world demonstrations. Emerald AI released results from a demonstration as part of EPRI’s DCFlex Initiative in Phoenix, Arizona, in collaboration with Oracle Cloud Infrastructure (OCI), NVIDIA, EPRI, and the regional power utility Salt River Project (SRP). And the demonstration showed that an AI compute cluster of GPUs in a commercial data center can reduce power consumption by 25% for 3 hours during critical periods of grid stress, such as a summer system peak load event, while maintaining satisfactory customer performance for AI workloads. Emerald AI is planning larger-scale demonstrations in Phoenix and across the country.
Emerald AI, participating in the NVIDIA Inception program for startups, is supported by other investors including BCG Chairman Rich Lesser, Google Chief Sustainability Officer Kate Brandt, Three Cairns co-founder Mark Gallogly, Galvanize Co-Chair Tom Steyer, Crusoe CEO Chase Lochmiller, Weights and Biases CEO Lukas Biewald, and others.
KEY QUOTES:
“We’re at a critical inflection point as exponential growth of AI computing pressures our electrical infrastructure. To unshackle AI technology progress from power constraints, Emerald AI transforms data centers from grid liabilities into flexible assets, enabling grid operators to swiftly interconnect AI, bolster reliability and energy security, and more efficiently harness the massive spare capacity on today’s grids.”
Emerald AI founder and CEO Dr. Varun Sivaram
“Emerald AI’s approach accelerates AI deployment while simultaneously enhancing grid reliability and shielding consumers from escalating energy costs. Enabling faster innovation without putting additional strain on our existing power systems is critical for maintaining America’s competitive edge in AI development.”
Former Secretary of Commerce and Emerald AI Advisor Gina Raimondo
“AI is driving unprecedented demand for energy. Without a solution to flexibly service that demand, innovation slows. Emerald AI’s technology can unlock up to 100 GW of untapped grid capacity by harnessing the infrastructure we already have. Emerald AI’s transformational software turns AI data centers from grid stressors to grid supporters.”
John Doerr, Kleiner Perkins Chairman
“The next 6-12 months will determine the gold standard for grid-responsive computing, and we’re confident Emerald AI is uniquely positioned to set that standard.”
Rob Toews, Radical Ventures Partner
