A self-driving laboratory that can accelerate materials discovery by a factor of 10 is now a reality and it’s also smarter and greener.
Researchers from North Carolina State University have unveiled a new AI-powered experimental system that uses real-time dynamic flow to collect continuous data transforming the traditional trial-and-error approach into a seamless, ever-learning process.
Self-driving labs are robotic platforms that combine machine learning and chemical automation to search for new materials. These labs aren’t sci-fi anymore they’re real tools that reduce research cycles from years to days, thanks to the ability to make data-driven decisions on the fly.
Traditionally, such labs used steady-state flow experiments, where a chemical reaction runs to completion before a single set of measurements is recorded. That means wasted time while the system waits for reactions to finish — sometimes as long as an hour per test.
Milad Abolhasani, the study’s lead author and ALCOA Professor of Chemical and Biomolecular Engineering said, “Instead of waiting around for each experiment to finish, our system is always running, always learning,” said Abolhasani.
The new method uses dynamic flow experiments, where chemical mixtures change continuously and are analyzed in real time capturing new data points every half second. For example, rather than a single readout at 10 seconds, the system now logs 20 distinct data points across that same time.
This shift means the lab’s AI can see how reactions evolve second by second, allowing for faster, more precise decision-making.
The team found their dynamic self-driving lab generated at least 10 times more data than conventional steady-state systems all while never stopping and never idling. More importantly, the system was able to zero in on the optimal material candidates on the first try after training, thanks to its rich, continuous data stream.
“The more high-quality experimental data the algorithm receives, the more accurate its predictions become,” explained Abolhasani.
Speed isn’t the only benefit. Because the AI gets smarter faster, it needs fewer experiments to reach a solution dramatically reducing chemical use and waste. That’s a major leap forward for environmentally responsible research, especially in fields like renewable energy, sustainable manufacturing, and green chemistry.
“Our approach means fewer chemicals, less waste, and faster solutions for society’s toughest challenges,” Abolhasani noted.
“The future of materials discovery is not just about how fast we can go — it’s also about how responsibly we get there.”
The study is published in Nature Chemical Engineering.