DOE Awards $34M to Accelerate Catalyst Discovery Using AI and Self-Driving Labs

The U.S. Department of Energy's Advanced Research Projects Agency-Energy (ARPA-E) has announced $34 million in funding for 12 projects that pair artificial intelligence with self-driving laboratory systems to dramatically accelerate industrial catalyst development. The announcement, made April 13, 2026, represents a significant step in applying AI to one of chemistry's most time-intensive challenges.

The funding flows through ARPA-E's CATALCHEM-E program — Catalytic Application Testing for Accelerated Learning Chemistries via High-throughput Experimentation and Modeling Efficiently — an initiative designed to compress catalyst development timelines from approximately ten years to roughly one year. The 12 selected projects will combine machine learning, AI-guided molecular design, and high-throughput experimentation to create continuous discovery workflows that can operate with minimal human intervention.

Notable recipients include the University of Wisconsin–Madison, which will apply $2.84 million toward catalysts for converting ethanol into specialty chemicals and fuels, and North Carolina State University, which received $2.99 million to develop catalysts that transform biomass and waste liquids into hydrogen-rich syngas. Ames National Laboratory will use $2.52 million to pursue precious-metal-free alternatives for hydrocarbon processing — a potential breakthrough for reducing industrial dependence on scarce and expensive platinum-group materials.

The initiative illustrates a broader trend of federal investment in AI-for-science applications, where AI models are used not just to analyze experimental data, but to actively direct physical laboratory operations. Autonomous labs equipped with robotic systems can test thousands of catalyst candidates per day under AI guidance, achieving in weeks what might take conventional research programs years.

For the broader AI industry, the ARPA-E announcement signals growing government confidence in deploying agentic AI in high-stakes scientific environments — a validation of the autonomous lab paradigm that several deep-tech startups have been building toward. The program also aligns with the DOE's broader $320 million Genesis Mission, which launched 26 AI-for-science grand challenges earlier this month.