NVIDIA AI Enhances Clean Energy Projects Across Five Companies
- Published
- Sep 21, 2026 — 10:00 UTC
NVIDIA AI Enhances Clean Energy Projects Across Five Companies
ThinkLabs AI has reduced the evaluation time for grid interconnection applications from 30-45 days to just 2 minutes for Southern California Edison. This significant improvement allows utilities to respond more rapidly to energy demands and uncertainties.
Atomic Canyon is leveraging AI for knowledge management in the nuclear sector, emphasizing that traditional nuclear processes need reinvigoration through AI to scale effectively. Trey Lauderdale, CEO of Atomic Canyon, noted that the current methods will not meet future energy demands.
PG&E operates the Diablo Canyon nuclear power plant, contributing to the 20% of U.S. electricity powered by nuclear energy. Meanwhile, Redwood Materials is developing off-grid power solutions specifically designed for AI factories, with CTO Colin Campbell stating that their systems can adapt to the fluctuations of AI training workloads.
TerraPower, co-founded by Bill Gates, is advancing its Natrium reactor, which aims to incorporate emissions avoidance as part of its energy solution. Chris Levesque, CEO of TerraPower, highlighted the company's vision established 20 years ago.
Commonwealth Fusion Systems (CFS) is working on fusion energy technology, with expectations to integrate fusion into the clean energy stack by the 2030s. Their SPARC tokamak design is 40x smaller than previous systems, while the ARC power plant is 10x smaller than traditional plants. Brandon Sorbom, CFS cofounder, emphasized the safety and cost-competitiveness of fusion energy compared to other sources.
These advancements illustrate how AI is being utilized to enhance efficiency and safety in clean energy production, marking a significant shift in energy management practices.
By Callan Zhang · Sep 21, 2026 · Editorial standards →
Summarised from the primary source with AI assistance under human editorial oversight. Turing Wire is not a primary source — read the original for the authoritative account.
Source: NVIDIA Blog
