Priya Donti's AI Tools Help Power Grids Integrate Wind and Solar — Cutting Costs and Emissions

AI-driven algorithms to make grids cleaner, cheaper and more reliable

Priya Donti develops machine-learning algorithms that help electricity networks manage variable wind and solar generation more effectively. A profile in MIT News traces how her personal experiences and academic path — from growing up in Massachusetts and traveling frequently to India to studying at Harvey Mudd College and earning a doctorate at Carnegie Mellon University — shaped her focus on climate impacts and equitable solutions.

From materials science to algorithms for the grid

Donti originally planned to study materials science and chemistry to improve solar cells, but exposure to computer science courses and research on using AI for power systems shifted her interests toward algorithms. She encountered work from U.K.-based researchers showing how machine learning can help electrical infrastructure handle intermittent renewable generation, which bridged her technical and environmental goals.

Practical tools for a real-world problem

The main challenge Donti addresses is operational: wind and solar output fluctuate, unlike steady coal or gas plants. Her models explicitly incorporate the physical constraints and scientific principles that govern power systems, producing better forecasts and management strategies that operators can use in everyday planning and control.

She developed an algorithm that runs faster than many existing methods while lowering operating costs. Unlike approaches that rely on crude approximations, Donti’s methods respect true physical limits of the grid, which has attracted interest from electrical network operators and utilities.

Synthetic data to expand research safely

Because real grid data is often restricted for competitive or security reasons, Donti produces realistic synthetic datasets so other researchers can build and test tools without accessing confidential information. These artificial datasets help accelerate research while protecting sensitive infrastructure details.

Impact and teaching

Improved network management can translate into lower household power bills and make it easier for clean energy sources to displace polluting generators. While completing her doctoral work at Carnegie Mellon, Donti helped launch Climate Change AI, a nonprofit that connects researchers, policymakers, and practitioners to share tools and resources for tackling climate problems.

Donti now teaches at MIT and will lead a spring course on applying AI to environmental challenges. She said she chose MIT in part for its collaborative culture and focus on societal impact:

"I knew that there would be an ecosystem of people who really cared, not just about success metrics like publications and citation counts, but about the impact of our work on society," she said.

Her work offers a practical path for integrating more wind and solar into grids while reducing costs and respecting system safety — a combination that could accelerate the transition to cleaner electricity.