Artificial intelligence may be part of the solution to the energy crisis it's creating. As AI systems consume more electricity, researchers are developing AI-powered tools to help the nation's electric grid modernize and operate more efficiently. At the U.S. Department of Energy's Genesis Mission Summit in July 2026, experts presented how AI foundation models could accelerate grid planning by up to 1,000 times, turning what once took months into minutes.

Why Is the Electric Grid Struggling to Keep Up With AI?

The electric grid is one of humanity's most complex systems, balancing electricity flow across vast networks while accounting for uncertain demand, weather, power generation, and new energy resources. Grid operators must plan not just for today but for hours, days, months, and years ahead. When all these variables combine, the number of possible future conditions can reach into the billions.

The problem is that traditional modeling approaches can only examine a small fraction of these scenarios. "Today we only analyze a few, which is crazy considering that we are in the midst of a massive and unprecedented expansion of our grid," explained Hendrik Hamann, a professor at Stony Brook University's School of Marine and Atmospheric Sciences and chief AI scientist for Innovation, Science and Security at Brookhaven National Laboratory.

"The electric grid is arguably the most important critical infrastructure, powering literally everything we depend on. The electric grid is also one of the largest and most complex systems humanity has ever built," said Hendrik Hamann.

Hendrik Hamann, Chief AI Scientist for Innovation, Science and Security at Brookhaven National Laboratory

With data centers and AI systems demanding unprecedented amounts of electricity, grid planners need better tools to evaluate whether and how major new electricity users can be integrated into the system. Without AI, these interconnection studies can take months or even years to complete.

How Can AI Foundation Models Transform Grid Planning?

Rather than running complex physics simulations for each scenario individually, researchers are developing grid foundation models that learn the relationships among the grid's physical structure, electricity demand, and power generation. These models are trained across millions of different grids and operating conditions, allowing them to rapidly predict how the system will behave under new circumstances.

Hamann described the approach as teaching AI to speak the grid's own language. "Unlike language foundation models, which are trained on text, our models are trained on the language of the grid. That means voltages, power, frequency across a broad set of conditions. So, the model learns how to predict the physics of the grid, enabling us to evaluate billions of scenarios rapidly, not only rapidly, but also accurately and robustly," he explained.

Hamann

"Without AI, data center interconnection simulation studies take months, sometimes years. Now we can do it literally in minutes. Without AI, we plan and operate the grid scenario-poor. But with AI, we can remove the blindfolds," stated Hendrik Hamann.

Hendrik Hamann, Chief AI Scientist for Innovation, Science and Security at Brookhaven National Laboratory

The results have already demonstrated remarkable potential. Some analyses have been accelerated by as much as 1,000 times, allowing grid planners to evaluate far more possible conditions, identify vulnerabilities, and make decisions based on a broader understanding of potential outcomes.

How to Implement AI-Driven Grid Solutions

Developing reliable grid foundation models requires extensive collaboration across multiple sectors. The effort brings together utilities, technology companies, universities, and national laboratories, each contributing data, infrastructure, and specialized expertise. Key implementation approaches include:

• Data Integration: Utilities must consolidate data from sensors, meters, modeling systems, and laboratories to create comprehensive datasets that AI models can learn from effectively.
• Collaborative Infrastructure: National laboratories, universities, and private companies work together to provide computing resources and domain expertise needed to train and deploy foundation models at scale.
• Open-Source Development: Hamann founded GridFM.org, a nonprofit community comprising more than 200 organizations and 500 members, to promote open-source collaboration in developing foundational AI technologies for the electric grid.

This collaborative approach is essential because grid modernization is not a single-company problem. It requires shared standards, interoperable systems, and contributions from utilities that operate different regions of the grid.

What Does This Mean for AI's Energy Problem?

The tension between AI's growing energy demands and its potential to strengthen the system supplying that energy is significant. AI is often viewed as a burden on the grid, but researchers argue the opposite can be true. By enabling faster, more comprehensive grid planning, AI can help utilities integrate new electricity demands more efficiently and identify ways to operate the system more sustainably.

The Genesis Mission, launched by the U.S. Department of Energy, represents a national commitment to harness AI for accelerating scientific discovery, strengthening energy systems, and enhancing national security. The inaugural summit brought together leaders from the Department of Energy, Congress, national laboratories, universities, and industry, with six inaugural awards involving Stony Brook researchers announced during the event.

As the nation faces unprecedented expansion of its electrical infrastructure to support both AI systems and the broader transition to clean energy, AI-powered grid planning tools may prove essential. The ability to evaluate billions of scenarios instead of a handful could mean the difference between a grid that can adapt to future demands and one that becomes a bottleneck for innovation and decarbonization.