NVIDIA and Emerald AI Join Leading Energy Companies to Pioneer Flexible AI Factories as Grid Assets
🤖 NVIDIA and Emerald AI are partnering with major energy companies including AES, Constellation, and NextEra Energy to pioneer a new class of flexible AI factories.
💡 These facilities will function as grid assets that can support the power grid during periods of stress rather than just being passive loads.
⚙️ The collaboration utilizes NVIDIA's Vera Rubin DSX AI Factory reference design, which includes the DSX Flex software library for connecting to power-grid services.
🔋 Factories can use co-located energy generation and storage as bridge power before transitioning to supplying the grid flexibly.
🧠 Emerald AI's Conductor platform will orchestrate computational flexibility alongside onsite generation and batteries to ensure quality service for AI compute tenants.
⚡ This architecture aims to unlock up to 100 gigawatts of capacity across the U.S. power system by optimizing infrastructure design.
🚀 The technology allows AI factories to connect faster by addressing the slow conventional interconnection timelines that hinder massive AI investment.
🏭 Energy partners like AES and Constellation emphasize that the current challenge is peak demand, not a total supply shortage of electricity.
💰 By producing valuable AI tokens and intelligence, these factories convert electricity into some of the highest-value outputs modern infrastructure can produce.
🛠️ The DSX Flex architecture enables operations to flex during limited periods of grid stress, reducing the need for broader grid expansion.
👁️ Jensen Huang stated that every system must be designed together—energy, compute, networking, and cooling—as a single architecture.
🌱 Joe Dominguez of Constellation noted that data centers have enormous potential for job creation, clean energy investment, and community benefits.
📉 The flexible operations approach helps avoid sizing infrastructure around peaks, which eases pressure on future system costs.
🔗 AES CEO Andrés Gluski highlighted that DSX Flex embeds flexibility from the outset, allowing AI infrastructure to operate as a grid asset.
🧩 The partnership demonstrates how cross-industry leaders can convene to support AI innovation while building a more reliable power system for Americans.
- NVIDIA and Emerald AI are collaborating with leading energy companies including AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power, and Vistra to pioneer flexible AI factories that connect to the grid faster.
- These new AI factories can harness co-located energy generation and storage as bridge power to bring AI capacity online faster while supporting broader value for customers and communities.
- The partnership projects could help unlock up to 100 gigawatts of capacity across the U.S. power system by optimizing infrastructure design and utilizing existing assets efficiently.
- Emerald AI's Conductor platform will orchestrate computational flexibility to shorten time on bridge power, support larger interconnections, and reduce infrastructure costs by easing pressure on future system peaks.
- Power-flexible AI factories can operate as grid assets that provide measurable relief during periods of grid stress, reducing the need for broader grid expansion to support reliability.
- NVIDIA's new Vera Rubin DSX AI Factory reference design includes DSX Flex software library specifically designed to connect AI factories to power-grid services for immediate performance and efficiency.
- Constellation noted that by using demand response capabilities from flexible AI factories, they can accommodate new load growth more efficiently while unlocking energy infrastructure investment and job creation.
- AES stated that their collaboration enables next-generation AI infrastructure to accelerate clients' time to power while allowing AI infrastructure to operate as a grid asset for faster, more efficient growth.
- The article suggests that 'conventional interconnection timelines can be too slow for the pace of AI investment,' forcing many gigawatt-scale projects to rely on co-located generation and storage as a stopgap measure.
- Permanently isolating energy resources from the grid carries significant risks, including leaving assets underutilized, raising long-term costs per AI token, and preventing those resources from contributing to overall grid reliability.
- The proposed solution relies heavily on 'bridge power' during the transition phase, which extends the duration where projects operate in a less efficient or integrated state before achieving full grid support.
- While the partnership aims to unlock 100 gigawatts of capacity, the text implies this is contingent on overcoming major hurdles such as infrastructure sizing around peaks and managing unprecedented demand spikes that existing systems struggle with.