Nvidia, Google and Emerald AI launch flexible data center consortium
Nvidia Corp., Google LLC and software startup Emerald AI Inc. have formed an industry consortium to promote the construction of flexible data centers. The AI Energy Management Alliance, or AEMA, officially launched on Wednesday. A flexible data center is a facility that cuts its power draw when the electrical grid is under strain. That usage […] The post Nvidia, Google and Emerald AI launch…
Three industry giants, Nvidia Corp., Google LLC, and software startup Emerald AI Inc., have joined forces to create the AI Energy Management Alliance (AEMA). The alliance aims to promote the construction of flexible data centers, which can reduce their power consumption when the electrical grid is under strain. These flexible data centers can adjust their energy usage by scheduling AI training runs during off-peak hours or by drawing power from on-site batteries.
The AEMA's three founding members are deeply involved in the flexible data center ecosystem. Google has developed demand response technologies that align its cloud facilities' energy usage with grid requirements. Nvidia offers a software tool called DSX Flex, which helps data center operators optimize their grid usage, on-premise energy generation equipment, and batteries.
Emerald AI, the third founding member, is a startup that raised $150 million in funding and is collaborating on a 100-megawatt data center in Manassas, Virginia, described as the "world's first power-flexible AI factory." Emerald AI's software platform, Emerald Conductor, enables server clusters to lower energy consumption during grid stress and minimize the impact on AI workloads.
The AEMA is also backed by 18 other companies, including Anthropic PBC, Analog Devices Inc., and multiple utilities. The consortium's primary goal is to develop best practices for flexible data centers, such as responding to power outages, and encourage data center builders to implement these standards before powering on new facilities.
Additionally, AEMA will standardize metrics for evaluating power optimization efforts and encourage industry players to share technical data.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.