Emerald AI's solution reduces power consumption for AI workloads, collaborating with tech giants and securing funding.

From NVIDIA: 2025-07-01 09:00:00

In many tech hubs, AI factories face long waits for power hookups due to energy infrastructure buildout. Emerald AI’s solution allows data centers to tap into existing energy resources more flexibly. Recent field tests in Phoenix show a 25% reduction in power consumption for AI workloads while maintaining performance quality. The startup secured over $24 million in seed funding and collaborated with NVIDIA, Oracle Cloud Infrastructure, and SRP for the Phoenix demonstration. Duke University estimates that AI data centers could unlock 100 gigawatts of new capacity by flexing their electricity consumption during peak demand periods. Emerald AI’s recent trial in the Oracle Cloud Phoenix Region showed a 25% reduction in power consumption compared to baseline load during peak demand. The International Energy Agency projects a doubling of electricity demand from data centers by 2030. Texas passed a law requiring data centers to reduce consumption or disconnect during load shed events. Emerald AI plans to expand its technology trials and continue working with NVIDIA to test its technology on AI factories.



Read more at NVIDIA: How AI Factories Can Help Relieve Grid Stress