Five Companies Using NVIDIA AI to Accelerate Clean Energy

Artificial intelligence is moving beyond chatbots and data centers and into one of the biggest infrastructure challenges facing the U.S.: how to produce, store and deliver more electricity with lower emissions.

At New York Climate Week, NVIDIA highlighted five companies using its computing platforms and software to tackle different parts of the clean-energy challenge. The companies are working on everything from smarter electrical-grid planning and nuclear operations to recycled battery storage and nuclear fusion.

The five companies are ThinkLabs AI, Atomic Canyon, Redwood Materials, TerraPower and Commonwealth Fusion Systems.

Their projects illustrate how AI is increasingly being used not simply to generate information, but to simulate complex physical systems, process large technical datasets and help engineers make faster decisions.

ThinkLabs AI Uses Digital Twins to Speed Up Grid Connections

One of the biggest obstacles to adding renewable power is connecting new generation projects to the electrical grid.

ThinkLabs AI is using digital twins and AI agents on the NVIDIA CUDA platform to simulate grid conditions and identify potential solutions to interconnection problems.

According to NVIDIA, Southern California Edison used ThinkLabs software to reduce the time required to evaluate an individual grid interconnection application from roughly 30 to 45 days to about two minutes.

The technology works by running grid simulations and helping identify potential barriers before new energy resources are connected.

ThinkLabs founder and CEO Josh Wong said the increasing uncertainty of the grid makes it important to evaluate conditions probabilistically rather than relying only on static assessments.

The potential benefit is significant: faster interconnection studies could help utilities evaluate new clean-energy projects without relying entirely on lengthy manual processes.

Atomic Canyon Brings AI Into Nuclear Power Operations

Atomic Canyon is focusing on another part of the energy system: nuclear power plants.

The company has developed AI-powered platforms called Neutron and NIVA, using NVIDIA accelerated computing to help nuclear professionals work with large amounts of technical and regulatory information.

Neutron is designed as an AI workbench for nuclear professionals. It can organize information including operating procedures, regulatory guidance, licensing records, design calculations and other structured and unstructured data into a searchable knowledge layer.

NIVA — the Nuclear Industry Virtual Assistant — was developed with the Institute of Nuclear Power Operations, the Electric Power Research Institute and the Nuclear Energy Institute.

The goal is to make it easier for nuclear operators and professionals to access complex information and use it in day-to-day operations.

Atomic Canyon CEO Trey Lauderdale argued that artificial intelligence could help modernize an industry that has relied on established processes for decades.

The technology is being positioned as a way to improve how existing nuclear infrastructure is operated rather than replacing the underlying nuclear technology.

Redwood Materials Uses Recycled EV Batteries for AI Power

The rapid expansion of AI is creating another energy problem: data centers need enormous amounts of electricity, while new grid infrastructure can take years to develop.

Redwood Materials is approaching the problem through battery storage.

The company is using recycled electric-vehicle batteries to build large-scale energy-storage systems that can provide power for AI infrastructure. NVIDIA says the systems can operate alongside clean-energy infrastructure and use an AI control layer running on its Blackwell platform.

The system is designed to respond to changing electricity requirements from data centers, including the fluctuations associated with AI workloads.

Redwood says its architecture can also reduce the need for some traditional electrical equipment, including transformers and inverters, while eliminating the need for a conventional uninterruptible power supply in its proposed configuration.

That could make it possible to deploy some power capacity faster while also giving retired EV batteries another use before their materials are eventually recycled.

The approach illustrates an increasingly important connection between two rapidly growing industries: AI and battery recycling.

TerraPower Uses Digital Twins for Advanced Nuclear Plants

TerraPower is taking a different approach by applying AI-powered digital-twin technology to advanced nuclear energy.

The company is developing its Natrium reactor, an advanced nuclear design that uses liquid sodium rather than water as its coolant.

According to NVIDIA, TerraPower is connecting an NVIDIA Omniverse-powered platform to digital-twin software intended to help accelerate the planning and delivery of future plants.

Digital twins can create virtual representations of physical systems, allowing engineers to model designs and processes before construction or deployment.

For a complex nuclear facility, the ability to test and visualize systems digitally could help engineers identify problems earlier and streamline parts of the development process.

TerraPower’s project is part of a broader effort to develop new nuclear technologies capable of providing electricity without direct carbon emissions during generation.

Commonwealth Fusion Systems Uses NVIDIA Technology for Fusion Research

The most futuristic project among the five comes from Commonwealth Fusion Systems, which is developing fusion energy.

CFS is building SPARC, a tokamak fusion machine in Massachusetts, while planning its larger ARC power plant in Chesterfield County, Virginia.

The company is using NVIDIA Omniverse libraries and OpenUSD to help accelerate engineering and experimentation around SPARC.

Fusion attempts to reproduce the basic process that powers the sun: combining light atomic nuclei under extreme conditions to release energy.

CFS uses high-temperature superconducting magnets to create strong magnetic fields while keeping its fusion systems relatively compact.

The company plans for its ARC facility in Virginia to produce about 400 megawatts of electricity when operational. CFS currently expects the plant to begin generating electricity in the early 2030s, subject to the necessary approvals and construction.

CFS has also taken concrete steps toward connecting ARC to the U.S. power system. In April 2026, the company announced that it had submitted an interconnection request to PJM, the large regional transmission organization serving Virginia and other states.

That makes the project more than a laboratory experiment: CFS is working toward integrating fusion generation into an actual commercial electricity market.

AI Is Becoming Part of the Energy Infrastructure

The five projects show several different ways artificial intelligence can be applied to energy.

ThinkLabs is using AI for grid planning and interconnection. Atomic Canyon is applying it to nuclear information and operations. Redwood Materials is combining AI with battery storage and power electronics. TerraPower is using digital twins for advanced nuclear development, while Commonwealth Fusion Systems is applying NVIDIA’s software ecosystem to fusion engineering and experimentation.

None of these technologies eliminates the physical challenges involved in building power infrastructure. Nuclear reactors, batteries, transmission systems and fusion machines still require major engineering, financing, regulation and construction.

But AI can potentially reduce the time needed for some of the information-heavy and simulation-heavy work surrounding those projects.

That matters as electricity demand rises and the power system faces pressure to add new generation, storage and transmission capacity.

For NVIDIA, the energy sector is also becoming increasingly important because the growth of AI itself is driving demand for more electricity.

The company’s clean-energy showcase therefore highlights a two-way relationship: AI needs more energy, while AI is increasingly being used to help build and manage the energy systems needed to supply it.

The next test will be whether these technologies can move from demonstrations and development projects into large-scale infrastructure that delivers measurable improvements in cost, speed, reliability and emissions.

What Happens Next

The five companies are at different stages of development.

ThinkLabs and Atomic Canyon are applying AI to existing energy operations and workflows. Redwood Materials is developing battery-based power systems for growing AI demand. TerraPower is advancing its Natrium technology, while CFS is moving toward its planned ARC fusion facility in Virginia.

For readers, the important development to watch is whether these AI-enabled projects can translate faster simulations and engineering workflows into actual power capacity connected to the grid.

That transition — from software and digital models to physical energy infrastructure — will determine how much impact AI can ultimately have on the clean-energy transition.