AI อะไรเนี่ย

Industry

NVIDIA Partners with Energy Leaders for Power-Flexible AI Factories

NVIDIA Partners with Energy Leaders for Power-Flexible AI Factories

What Happened: AI Factories as Dynamic Grid Assets

Last week at CERAWeek, a prominent gathering for energy leaders, NVIDIA and Emerald AI introduced a groundbreaking approach to integrating AI factories into our energy infrastructure. Instead of viewing AI factories as static power consumers, they proposed treating them as flexible, intelligent assets that can dynamically interact with the power grid.

This innovative concept is built upon the robust NVIDIA Vera Rubin DSX AI Factory reference design combined with Emerald AI’s Conductor platform. This synergy brings together accelerated computing, AI factory architectures, and real-time energy orchestration, allowing large AI deployments to connect faster, operate more efficiently, and bolster grid reliability. The collaboration underscores a significant shift in how we power the future of AI. You can read more about this partnership in NVIDIA's official announcement: NVIDIA and Emerald AI Partner for Flexible AI Factories.

Why It Matters: Fortifying the Grid and Boosting Efficiency

The rapid expansion of AI necessitates a re-evaluation of energy consumption. This new approach by NVIDIA and Emerald AI enables AI factories to generate high-value AI tokens while intelligently responding to grid conditions – flexing their power demand when needed, supporting overall grid stability, and reducing the need for excessive infrastructure buildout to handle peak loads.

A consortium of major energy companies, including AES, Constellation, Invenergy, NextEra Energy, Nscale Energy & Power, and Vistra, are actively collaborating to build the necessary energy generation capacity. These companies plan to develop optimized generation strategies, including hybrid projects with co-located power sources, specifically designed to support AI factories built on the NVIDIA and Emerald AI architecture. This partnership ultimately strengthens grid reliability by pairing large AI loads with flexible operations, new generation resources, and intelligent controls.

NVIDIA has a long-standing commitment to energy efficiency. Impressively, from the NVIDIA Kepler GPU in 2012 to the current NVIDIA Vera Rubin platform, the number of tokens generated within the same power budget has increased by over 1 million times. This focus on "tokens per second per watt" is crucial for sustainable AI scaling, as described in NVIDIA founder and CEO Jensen Huang's concept of a five-layer AI cake, where energy forms the foundational layer.

Accelerating the Future: AI, Digital Twins, and Workforce Development

Beyond the core partnership, NVIDIA's ecosystem partners showcased a range of advancements at CERAWeek, demonstrating how AI, simulation, and workforce innovation are accelerating the energy infrastructure required for the intelligence era.

  • Robotics in Solar: Maximo, a solar robotics company incubated at AES, announced the completion of a 100-megawatt robotic solar installation at AES’ Bellefield site. This utility-scale deployment, using AI-driven robotics developed with NVIDIA accelerated computing, NVIDIA Omniverse libraries, and the NVIDIA Isaac Sim framework, showcases how autonomous installations can improve speed, safety, and consistency.
  • Digital Twins for Nuclear Energy: TerraPower, in collaboration with SoftServe, previewed an NVIDIA Omniverse-powered digital twin platform. This platform is designed to drastically shorten advanced nuclear plant siting and design timelines from years to mere months by applying AI and simulation to early-stage engineering. This kind of blueprint is essential for the future of AI factories, leveraging the power of NVIDIA Omniverse DSX Blueprint for AI Factories.
  • Skilled Workforce for AI Infrastructure: Adaptive Construction Solutions launched a national registered apprenticeship initiative, in partnership with NVIDIA, to cultivate the skilled workforce needed for AI factories and energy infrastructure. This program aims to scale training for critical trades, expanding access to high-demand careers.

The Path Forward: Scaling AI Factories for Grid Reliability

GE Vernova, Schneider Electric, and Vertiv also highlighted their contributions, emphasizing how digital twins, validated reference designs, and converged infrastructure are becoming indispensable for scaling AI factories as reliable grid participants. These efforts directly address the "power-to-rack" challenge, ensuring that AI infrastructure is designed from the outset as an integrated energy and compute system.

By integrating AI factories as flexible assets into the energy grid, these collaborations pave the way for a more resilient, efficient, and sustainable future for both AI and global energy systems.

Read more:

Explore the full details of this initiative and NVIDIA's vision for energy-efficient AI factories: Efficiency at Scale: NVIDIA, Energy Leaders Accelerating Power-Flexible AI Factories to Fortify the Grid.