Nvidia and Emerald AI are working with AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power and Vistra to power and advance a new class of AI factories that connect to the grid faster, generate valuable AI tokens and intelligence, and operate as flexible energy assets that can support the grid.
By bringing together technology, energy and infrastructure leaders, the collaboration demonstrates how companies across industries can convene to support AI innovation in the United States, while building a more reliable power system for Americans.
These next-generation AI factories will harness the new Nvidia Vera Rubin DSX AI Factory reference design, which includes the DSX Flex software library for connecting AI factories to power-grid services.
For accelerated deployment, the factories can use co-located energy generation and storage as bridge power for hybrid AI factories, then later harness these resources to flexibly supply the grid, accelerate AI factory interconnection and support the broader power system. This approach helps bring AI capacity online faster while creating broader value for customers and communities.
The DSX reference architecture can also support flexible AI factories without co-located energy resources to achieve larger and faster power grid connections.
Emerald AI’s Conductor platform will orchestrate computational flexibility alongside onsite generation, batteries and other behind-the-meter resources to deliver precise, grid-responsive power flexibility while ensuring quality of service for AI compute tenants. This coordination helps operators meet power targets, protect priority workloads, shorten time on bridge power, and support larger and faster interconnections. It can also help reduce the need for infrastructure to be sized around peaks, easing pressure on future system costs.
“AI factories are the engines of the intelligence era, and like any great engine, every system must be designed together — energy, compute, networking and cooling as one architecture,” says Jensen Huang, founder and CEO of Nvidia.
“Nvidia and Emerald AI are working together to enable a future for AI where performance, efficiency and grid responsiveness can be tapped into immediately.”
Varun Sivaram, founder and CEO of Emerald AI, comments: “AI factories are too valuable to be treated as either passive loads or permanent islands. They produce tremendously valuable AI tokens and knowledge, and with DSX Flex, they can also provide measurable relief back to the grid.
“Emerald Conductor orchestrates compute flexibility alongside onsite energy resources to support the grid, so projects can connect sooner, preserve quality of service for AI tenants and ultimately strengthen the power system around them.”
Building AI factories that strengthen the grid
Today’s electric systems are built to serve peak demand but are underutilised during most hours of the day.
Power-flexible AI factories can help unlock up to 100 gigawatts of capacity across the US power system by combining optimised infrastructure design with efficient use of existing assets and, where needed, new-build generation, while flexing during limited periods of grid stress to reduce the need for broader grid expansion to support reliability.
AI factories convert electricity into AI tokens, models and intelligence — among the highest-value outputs modern infrastructure can produce. Meeting that opportunity will require innovation in computing as well as in how companies plan, build and operate energy infrastructure.
Many gigawatt-scale AI projects are turning to co-located generation and storage because conventional interconnection timelines can be too slow for the pace of AI investment.
However, permanently isolating generation and storage from the grid has drawbacks. It can leave assets underutilised, raise long-term cost per AI token and prevent energy resources from supporting grid reliability.
AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power and Vistra aim to build the energy generation capabilities necessary to ensure supply meets surging demand.
The companies will collaborate to evaluate optimised generation applications designed to power the AI factories built with the architecture developed by Nvidia and Emerald AI, including through hybrid projects that use co-located power, to speed time to power and create value for the broader grid.
By pairing large AI loads with flexible operations, new energy generation capabilities and intelligent controls, this approach can help boost grid reliability.
The companies can also support flexible AI factories that are grid-connected from the outset, using co-located energy resources if available.