Emerald AI, Google, and Nvidia have announced the launch of the AI Energy Management Alliance (AEMA) – a coalition aimed at accelerating the connection of data centers for artificial intelligence to the energy grids of the United States. The initiative involves creating infrastructure capable of dynamically adjusting electricity consumption based on network load, which should simplify the scaling of AI computations and reduce pressure on the energy system.
AEMA will bring together participants from the entire AI and energy ecosystem – from platform developers and data center operators to electricity producers, technology companies, utilities, and regional network operators. Participants will focus on creating unified technical and operational approaches, collaborating with energy companies, and promoting regulations for data centers capable of responding to network conditions. In particular, a flexible data center can shift computational loads, utilize energy storage, engage backup generation, temporarily reduce electricity consumption, and respond to emergencies in the energy system. AEMA plans to evaluate such facilities based on measurable metrics – response speed, duration, predictability, and behavior during emergencies, rather than on specific equipment or software.
The demand for new computing power is already driving companies to seek alternative approaches to placing AI infrastructure. In May of this year, Google and SpaceX began negotiations about placing computing power in space. Google plans to launch the first prototypes of satellites by 2027, while SpaceX considers orbital data centers as a potential alternative to ground infrastructure.
Related articles
