AI Energy Management Alliance by Emerald AI, Google and NVIDIA

Emerald AI, Google, and NVIDIA have established the AI Energy Management Alliance (AEMA), an innovative collaboration aimed at developing data centers that can dynamically adjust their electricity consumption to the conditions of the electricity grid. This flexibility will not only support the growth of AI infrastructure but also contribute to more efficient use of existing energy capacity and reduce environmental impact.

The AEMA aims to build an infrastructure that is not only connected to the grid but actively collaborates with it. Traditional network connection processes were not designed for the demands of modern computing infrastructure that must be capable of intelligently responding to limited energy resources. A flexible data center can adjust its electricity consumption in various ways, such as shifting computing loads or utilizing storage capacity.

Furthermore, AEMA is technology-neutral and focuses on performance-based requirements. Emphasis is placed on measurable services that a facility can provide, such as responsiveness, predictability, and operational reliability in emergencies. The alliance aims to promote standardization and transparency, giving developers and grid operators more certainty in the performance of data centers and their effects on the grid.

With the establishment of AEMA, a wide range of stakeholders comes together, including AI platforms, infrastructure providers, and energy suppliers. By developing joint technical and operational approaches and collaborating with utilities, AEMA aims to promote the integration of AI facilities into the electricity grid. This will accelerate the growth of AI infrastructure in the US and ensure that the necessary grid is built quickly and sustainably.

Read the full article from Nvidia.