NVIDIA Highlights AI’s Role in Sustainable Energy Transition at Climate Week NYC, Showcasing Advances in Energy Efficiency, Carbon Reduction, and Climate-Resilient Infrastructure

At Climate Week NYC, NVIDIA highlighted the pivotal role of accelerated computing in driving the sustainable energy transition. The company emphasized that energy efficiency in large language model inference has improved by 100,000 times over the past decade, proving that advanced computing is inherently sustainable. The event, held through September 26 in New York City, unites experts from research, industry, government, and nonprofits to tackle climate challenges. This year’s focus is on energy—its sources, scalability, and how AI can optimize power grids in transformative ways. NVIDIA showcased its climate technologies and released two new product carbon footprint reports detailing the emissions intensity of its GPUs. These reports underscore the company’s commitment to transparency and sustainability in hardware development. AI is not inherently energy-intensive—it can be a powerful tool for sustainability. By rapidly detecting anomalies in energy systems, AI enables faster responses to grid issues, preventing larger disruptions. According to projections, widespread adoption of AI across key sectors could save nearly 4.5% of total energy demand by 2035. Projected AI-Induced Energy Savings by 2035 Across Energy-Intensive Sectors: Industry: Iron and Steel – 31,160 petajoules Industry: Cement – 4,500 petajoules Industry: Chemicals – 210,440 petajoules Industry: Aluminum – 4,260 petajoules Industry: Paper – 21,860 petajoules Industry: Other – 813,650 petajoules Transportation: Light commercial vehicles – 68,160 petajoules Transportation: Heavy duty trucks – 33,670 petajoules Transportation: Cars – 34,460 petajoules Transportation: Buses – 61,690 petajoules Transportation: Aviation – 43,120 petajoules Transportation: Shipping – 4,940 petajoules Transportation: Rail – 7,530 petajoules Buildings: Residential – 14,780 petajoules Buildings: Non-residential – 41,760 petajoules These figures are based on analysis from the International Energy Agency and Princeton University’s Net-Zero America Project, as cited in a CSIS report on AI and energy. During a Climate Week panel, NVIDIA joined Crusoe Energy Systems and Emerald AI in discussing how AI can scale clean energy infrastructure and improve grid reliability. Emerald AI, a startup in NVIDIA’s Inception program and part of its Sustainable Futures initiative, is developing AI tools to manage data center power use during peak grid demand. Emerald AI is also collaborating with NVIDIA on a new Omniverse Blueprint for building AI infrastructure that is energy-efficient and grid-friendly. This reference design turns data centers into intelligent, power-flexible AI factories where every watt contributes directly to computing performance. “By helping build AI factories that are power-flexible, we’re unlocking 100 gigawatts of unused grid capacity and solving AI’s energy bottleneck,” said Varun Sivaram, CEO of Emerald AI. “This is a game-changer for clean, reliable, and affordable energy.” NVIDIA continues to reduce its own environmental impact. Its latest product carbon footprint report shows a 24% reduction in embodied carbon emissions intensity between the HGX H100 and HGX B200 baseboards. The company plans to publish similar reports for future products. All NVIDIA offices and data centers under its control run on 100% renewable energy. For leased data centers, the company purchases carbon-free electricity to cover 100% of their footprint. NVIDIA Earth-2, the company’s platform for simulating weather and climate at scale, is being used to improve energy system resilience. High-resolution AI models help utilities anticipate storm damage, optimize maintenance, and better predict renewable energy output from wind and solar sources. These tools are featured across multiple Climate Week panels, including those hosted by Columbia University, Google, and AWS, demonstrating how AI can accelerate climate action and support a decarbonized future.