US awards $500M to Alphabet spinoff for AI-driven chip materials

The U.S. government is investing $500 million in an Alphabet spinoff to accelerate the development of semiconductor materials using artificial intelligence. SandboxAQ, a company focused on AI and quantum technologies, announced on 17 June 2026 that it had secured the funding under the CHIPS and Science Act to address critical gaps in the domestic semiconductor supply chain. The grant will support research into new materials for chip manufacturing, including alternatives to PFAS chemicals, catalysts, rare-earth-free magnets, and batteries less dependent on foreign-sourced lithium and other minerals.

How the funding will be used

SandboxAQ will direct the $500 million toward four key areas: PFAS-free semiconductor materials, new fabrication catalysts, rare-earth-free magnets, and batteries that minimize reliance on foreign-sourced minerals. The company’s approach relies on LQMs, which generate material predictions for laboratory testing. Unlike traditional AI models trained on human language, these systems are designed to adhere to the laws of physics and chemistry, though they also incorporate synthetic data where real-world experimental data is lacking.

A SandboxAQ spokesperson told The Register that the company already uses real-world data where available but can proceed with synthetic data when necessary. The spokesperson acknowledged the risk of compounding errors in AI-driven simulations but emphasized that lab validation serves as a final checkpoint. "A material either performs in the lab, or it doesn’t," the spokesperson said, adding that this process prevents unrealistic predictions from derailing research.

SandboxAQ has previously worked on catalysts, battery materials, alloy discovery, and PFAS mitigation, and these efforts will inform its CHIPS Act-funded projects. The company claims its AI-driven workflows have already reduced candidate screening timelines from months to weeks in commercial deployments. However, it cautioned that semiconductor industry qualification processes remain rigorous and time-consuming, even if the path to adoption does not require building new fabrication facilities from scratch.

Why the project matters for the semiconductor industry

The U.S. semiconductor industry has long struggled with supply chain vulnerabilities, particularly in sourcing critical materials like rare earths and PFAS chemicals. Over 80% of rare earth elements, essential for magnets in chip manufacturing, are currently mined and processed in China. Similarly, PFAS chemicals, widely used in semiconductor fabrication, face growing regulatory scrutiny due to their environmental persistence and health risks. The CHIPS Act funding aims to mitigate these dependencies by fostering domestic alternatives.

SandboxAQ’s AI-driven approach mirrors efforts in pharmaceutical research, where AI has been used to accelerate drug discovery. However, the success of AI in materials science remains unproven at scale. The U.S. National Institutes of Health noted in 2025 that no AI-designed drug had yet reached functional approval, raising questions about whether AI can deliver on its promises in semiconductor materials. SandboxAQ’s reliance on synthetic data and physics-based models may address some of these concerns, but the project’s outcomes will depend on real-world validation.

What to watch

The $500 million grant is part of a broader $52 billion CHIPS Act initiative, which has already seen mixed results. While the U.S. government’s 10% stake in Intel in 2024 helped stabilize the company, domestic semiconductor manufacturing still lags behind global competitors. SandboxAQ’s project could take years to yield commercially viable materials, and its success will hinge on collaboration with existing fabs and equipment suppliers. Industry observers will be watching for updates on lab validation, partnerships with manufacturers, and regulatory approvals for any new materials developed under the program.