NVIDIA, Emerald AI Partner on Flexible AI Factories as Grid Assets

NVIDIA and Emerald AI announced that they are partnering with six leading U.S. energy companies AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power and Vistra to develop a new generation of AI data centers designed to connect to the power grid faster and operate as flexible energy assets.

NVIDIA and Emerald AI on Sunday announced a broad industry collaboration to develop “flexible AI factories”, a new class of artificial intelligence data centers engineered to accelerate grid connections, generate AI tokens and intelligence, and function as dynamic assets that can support the wider power grid.

The announcement was made at CERAWeek 2026 in Houston. Joining the effort are six major U.S. energy companies: AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power and Vistra.

The partnership combines NVIDIA’s new Vera Rubin DSX AI Factory reference design which includes the DSX Flex software library for connecting AI factories to power-grid services with Emerald AI’s Conductor platform, which orchestrates computational flexibility alongside onsite generation, batteries and other behind-the-meter energy resources.

Traditional data centers draw a fixed, high level of power around the clock. The new flexible AI factory model takes a different approach, facilities can ramp electricity consumption up or down in response to grid conditions, in a manner similar to industrial demand-response programs, rather than running at constant full capacity.

For faster initial deployment, factories can use co-located energy generation and storage as bridge power, then later redirect those resources to support the broader grid. The DSX architecture also supports AI factories that connect directly to the grid without co-located generation, enabling larger and faster interconnection agreements.

Emerald AI’s Conductor platform will coordinate compute workloads, onsite generation and battery storage to deliver precise, grid-responsive power flexibility. The company says this approach can help operators meet power targets, protect priority workloads and shorten the time facilities spend on bridge power ahead of full grid connection.

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. NVIDIA and Emerald AI are working together to enable a future for AI where performance, efficiency and grid responsiveness can be tapped into immediately.
Jensen Huang, founder and CEO, NVIDIA

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.
Varun Sivaram, founder and CEO, Emerald AI

Today’s electric systems are designed to meet peak demand but remain underutilized for most hours of the day. The companies say power-flexible AI factories could help unlock up to 100 gigawatts of capacity across the U.S. power system by combining optimized infrastructure design with efficient use of existing assets and by flexing power consumption during limited periods of grid stress.

Many large-scale AI projects have turned to co-located generation and storage because conventional interconnection timelines can be too slow to keep pace with AI investment. However, permanently isolating those resources from the grid can leave assets underutilized, raise the long-term cost per AI token and prevent generation from contributing to grid reliability, the companies said.

The six energy partners AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power and Vistra have committed to building energy generation capabilities to match surging AI demand, and will collaborate to evaluate generation applications optimized for the new AI factory architecture.

The companies plan to deploy DSX Flex at NVIDIA’s AI Factory Research Center in Virginia later in 2026. A separate facility, the Aurora AI Factory being built by Digital Realty in Manassas, Va. is slated to open in the first half of 2026 and will serve as the first facility built to the new flexible AI factory standard.

AI data centers have emerged as one of the fastest-growing sources of electricity demand in the United States, prompting utilities and grid operators to grapple with how to expand transmission and generation capacity quickly enough to meet the load. The NVIDIA-Emerald AI initiative represents an effort to reframe the problem: rather than treating AI data centers purely as power consumers, advocates argue they can be redesigned to contribute to grid stability.

The announcement at CERAWeek, the premier annual energy industry conference signals a growing convergence between the AI and energy sectors, as companies seek solutions that can simultaneously advance computing capability and grid reliability.

Disclaimer: This article is based on information from NVIDIA’s official announcement and publicly available sources. NervNow has not independently verified the claims.

MORE ON AI & INFRASTRUCTURE:
EdgeConneX Starts Work on 200MW AI Data Center in Osaka
S&P Global Buys Enertel AI for Power Price Forecasting
With Sarvam and Krutrim, Has ‘Make in India’ in AI Finally Arrived?
Oasis Security Raises $120 Million to Govern AI Agent Access
Circles, Huawei Sign Deal to Build AI-Native Telecom Solutions