Emerald AI Has a New Approach to Meeting AI’s Energy Demand

President Donald Trump and members of his cabinet recently highlighted the growing energy demand needed to power the expansion of artificial intelligence (AI) at an event in Pennsylvania. Republican Senator Dave McCormick organized an AI and energy summit at Carnegie Mellon University in Pittsburgh, where he and other leaders emphasized the importance of developing AI for economic and national security.

“This is a competition we must win,” said McCormick during his opening remarks.

Billionaire businessman Jonathan Gray, president and CEO of Blackstone Group, noted that access to energy has become a major constraint on AI development. Energy analysts predict a significant spike in electricity demand due to the energy consumption of data centers. A Department of Energy report projected that U.S. data center electricity use could nearly triple over three years, potentially consuming 12 percent of the country’s total electricity by 2028.

Tech companies are investing heavily in new energy sources and securing their own power supplies. Grid managers and utility companies are grappling with how to meet this growing demand while maintaining affordable and reliable power supply.

Sustainability is also at risk as AI growth drives up greenhouse gas emissions, causing many tech companies to fall behind on their net-zero climate goals.

Emerald AI is developing technology that makes data centers more flexible in their power consumption. “Our goal is to make these data centers flexible,” Varun Sivaram, CEO and founder of Emerald AI, told Newsweek. Emerald’s platform allows grid managers to remotely shift power demand for data centers without affecting AI performance. This can connect more data centers to today’s energy system and use existing infrastructure more efficiently.

Emerald recently secured $24.5 million in financing from tech, energy, and climate leaders, including Nvidia, former climate envoy John Kerry, and Kleiner Perkins chair John Doerr.

Sivaram explained that Emerald’s technology enables power managers to reduce AI chip clusters’ power consumption during peak demand periods without impacting performance. In a recent test in Phoenix, Emerald reduced AI chip cluster power by 25 percent for three hours when the city needed most energy.

“This approach can turn data centers into grid allies instead of a potential threat,” said Sivaram. “Communities fear rate increases and blackouts due to dirty diesel generators used by some data centers. Emerald’s flexibility mitigates these concerns.”