SandboxAQ Uses AI to Purge 'Forever Chemicals' From Chip Manufacturing

SandboxAQ is deploying artificial intelligence to eliminate per- and polyfluoroalkyl substances (PFAS) from the semiconductor supply chain. Backed by a $500 million research and development award from the U.S. Department of Commerce, the initiative aims to replace these toxic materials without compromising chip performance.

Funded through the CHIPS and Science Act, the project utilizes SandboxAQ’s ReAQT platform and Large Quantitative Models (LQMs) to screen millions of candidate materials. These AI-driven tools identify alternatives to PFAS, which are currently ubiquitous in chip fabrication due to their electrical insulation and heat resistance. To establish the technical foundation for this search, SandboxAQ previously partnered with AWS, Intel, and Accenture, utilizing over 1.1 million vCPUs to simulate the bond-breaking chemistry of toxic PFAS molecules with near-exact precision.

The PFAS Dilemma

PFAS, commonly known as "forever chemicals," are prized in the semiconductor industry for their extreme chemical stability and resistance to heat. However, these same properties make them environmental hazards, as they persist indefinitely in nature and are linked to significant toxicity. For the U.S. government, the reliance on these chemicals represents both an environmental liability and a national security risk; discovering domestic, sustainable alternatives is essential for supply chain independence.

Accelerating Discovery

Replacing PFAS is a formidable technical challenge because any substitute must meet the same rigorous performance standards as the original chemicals. Traditional materials science relies on trial-and-error laboratory work, a process that can take years to yield results. By shifting to quantum-inspired simulations and LQMs, SandboxAQ intends to compress these development timelines from years to days.

This investment in AI-enabled discovery advances the ability to identify novel chemistries and molecules for the semiconductor ecosystem while improving U.S. supply chain resilience. By removing a major environmental hazard and a potential supply chain vulnerability, the project seeks to stabilize the domestic chip industry.

Future Outlook

As the project progresses, the industry will watch whether these AI-screened materials can transition from simulation to high-volume manufacturing. While computational screening identifies the most promising candidates, final verification requires physical integration into the fabrication process. The success of this effort could provide a blueprint for using Large Quantitative Models to solve other critical material shortages across the broader electronics sector.