Nvidia (NASDAQ: NVDA, GSE: NVD) announced on Monday that it is working with Emerald AI to develop a new type of "artificial intelligence factory" with a number of major US energy companies ". Such facilities are designed to operate as flexible assets on the power network and help accelerate the deployment of large-scale computing infrastructure.
The move was unveiled at a CERAWeek conference in Houston. Participants include AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power, and Vistra.
The parties plan to combine AI infrastructure design with power generation and grid integration strategies to reduce the time it takes for data center capacity to come online and support grid reliability. The proposed "artificial intelligence factory" will be built using Nvidia's Vera Rubin DSX reference architecture, which includes DSX Flex software. The software is designed to enable facilities to adjust power consumption based on grid conditions.
Emerald AI Conductor platform will coordinate computing with on-site energy resources (such as power generation and battery storage), helping to balance power demand while maintaining the performance of artificial intelligence computing. The companies say the model is designed to cope with the growing demand for electricity from AI workloads. Supporting power generation and storage facilities allows the facility to access power more quickly and then connect deeper to the grid.
During peak demand, these resources can also be used to reverse power to the grid. Unlike the traditional fully off-grid data center model, this approach is designed to keep energy assets connected to the system and respond to the needs of the broader system. According to the companies, the move could increase utilization of existing infrastructure and reduce the need for grid expansion by allowing AI facilities to regulate consumption up and down during periods of stress.
The collaboration also reflects a growing focus on combining large data center loads with flexible energy strategies, as traditional grid interconnection schedules have struggled to keep pace with AI-related investments. Some projects have shifted to dedicated power generation to speed deployment, but without grid integration, this could limit broader system benefits.
Participating energy providers said they will evaluate power generation and infrastructure configurations to support the AI factory model, including hybrid setups that combine grid access with on-site power. The company pointed out that the combination of flexible operation and new generation capacity can help to meet the rising power demand while maintaining system reliability.
Nvidia and Emerald AI say they have tested some elements of the concept in five commercial data centers around the world in the past year. Commercial-scale deployment of the DSX Flex system is expected to begin later this year at an NVIDIA AI research center in Virginia.
The partners said they plan to identify and use the architecture to advance more projects, focusing on accelerating AI infrastructure deployment, improving grid integration, and supporting broader energy system capabilities. "The AI factory is the engine of the intelligent age, and like any great engine, every system must be designed in one piece-energy, computing, networking and cooling as a single architecture," said Renxun Huang, CEO of Nvidia.
"NVIDIA and the Emerald AI are working together to unlock an AI future where performance, efficiency and grid responsiveness make an immediate difference." Affected by the news, Nvidia shares rose nearly 3%.