SandboxAQ has signed a definitive agreement with the US Department of Commerce’s CHIPS Research and Development Office for a $500-million award to address one of the most urgent challenges in American manufacturing — the foreign control of critical materials and chemistries that are essential to semiconductor manufacturing.
The award provides funding to develop novel molecules and formulations for semiconductor manufacturing within four programmatic areas: PFAS-free process chemicals, catalysts, rare earth-free magnets, and battery systems.
SandboxAQ will then advance the strongest breakthrough results into scaled domestic manufacturing and commercialization, via high-performing American manufacturing partners. This funding supports R&D across target categories in which foreign supply suppressed domestic production for decades and will ultimately strengthen national and economic security.
SandboxAQ will invest in enhancements to its ReAQT software platform and Large Quantitative Models (LQMs) to accelerate its work in virtually screening millions of candidate materials, after which it will select the most promising to validate with lab partners.
LQMs are AI systems trained on the laws of physics, chemistry, and biology, not human language. What otherwise would take decades of laboratory trial-and-error can now run as a targeted, AI-driven campaign.
The award allocates the funding across four material programmatic areas and for foundational investment in SandboxAQ’s core LQM platforms for advanced chemical and materials development critical to semiconductor manufacturing. In connection with the award, the US Department of Commerce will receive a minority, non-voting equity stake in SandboxAQ.
“This award will accelerate the discovery and innovation of critical materials and reduce our reliance on foreign-controlled materials,” said Secretary of Commerce Howard Lutnick.
Jack Hidary, CEO of SandboxAQ, said securing America’s semiconductor future means controlling the materials that drive this vital sector.
“SandboxAQ’s large quantitative models are grounded in the engineering and physics needed to address the needs of our domestic semiconductor sector. This award from the US Department of Commerce enables SandboxAQ to run advanced AI-driven programs across four critical material categories and then work with partners to scale the resulting formulations,” added Hidary.
The four programmatic areas of the award include, first, per- and polyfluoroalkyl substances (PFAS) which are “forever chemicals” that appear throughout chip manufacturing as heat-transfer fluids, lubricants, insulating coatings, and surface treatment chemicals, and no compliant alternatives yet exist at scale.
US semiconductor factories that cannot certify PFAS-free alternatives may risk simultaneous supply disruption and regulatory exposure that could force production cutbacks in newly built domestic facilities. SandboxAQ has developed approaches to PFAS breakdown to address this issue and will build on this work with the CHIPS Act award.
Second, SandboxAQ will build on the progress already made by its AQCat workflows to screen catalyst candidates at near-quantum-chemistry accuracy 20,000 times faster than traditional methods, and reduce catalyst development timelines in commercial deployment.
Third, SandboxAQ will use ReAQT and its LQMs to screen magnet chemistries that eliminate or sharply reduce reliance on neodymium and other heavy rare earth elements, at a speed and precision no prior method has matched, targeting formulations that can be manufactured using existing US production equipment lowering the capital barrier to commercialization.
And fourth, most chip factory backup power systems depend on battery materials (lithium, cobalt, key chemical precursors) that are heavily concentrated overseas. As a result, a geopolitical or supply chain shock could potentially disrupt a US semiconductor factory.
Thus, SandboxAQ will build on the progress already made by its AQVolt workflows, which is a frontier AI model for battery chemistry. This programmatic area will develop battery chemistries that do not depend on lithium and other materials that have foreign chokepoints.
One platform across the four areas
ReAQT, SandboxAQ’s AI simulation platform, is the foundation for all four material programmatic areas and is built to operate at scale. SandboxAQ plans to deepen its investment in ReAQT in order to accelerate the development timeline for new materials discovery.
ReAQT generates its own physics-grounded training data through high-fidelity simulations, including Density Functional Theory, Molecular Dynamics, and reaction modeling, then trains SandboxAQ’s proprietary Large Quantitative Models (LQMs) on that data and integrates them directly into Design-Make-Test workflows.
Because LQMs learn from physical laws and real world data, they deliver accurate predictions about materials that have not been previously synthesized, giving researchers a reliable map of what is possible before committing to expensive lab work.
”The most accurate simulation methods are too slow to search the range of materials that matter at scale. Models trained purely on existing data are fast but break down when applied to materials they have never seen,” said Stefan Leichenauer, VP of engineering at SandboxAQ.
“ReAQT solves both problems by generating its own high-quality training data grounded in physics, then training our Large Quantitative Models on it. The result is a platform that makes reliable predictions about materials, compressing development timelines in ways that shift what is commercially viable,” said Leichenauer.
