SandboxAQ receives a $500 million award from the US to identify chip materials.

      The US government has recently invested in an AI startup, wagering that artificial intelligence can help reduce China’s dominance over the materials used in every chip manufacturing facility. According to Reuters, the Department of Commerce has granted SandboxAQ $500 million under the CHIPS Act to create new chemicals and metals for domestic semiconductor production. In exchange, Commerce acquires a minority, non-voting equity stake in the company and potential royalties if the project succeeds.

      SandboxAQ, which has the support of Nvidia and was valued at $5.75 billion last year, aims to develop four materials that are currently reliant on foreign supplies for US chipmaking: substitutes for the “forever chemical” PFAS, catalysts, magnets that do not utilize Chinese rare earths, and batteries that do not require imported lithium.

      The approach taken here is significant. This arrangement is more than just a straightforward grant. By obtaining equity and a share of future royalties, the government is acting like a venture capital fund, following a similar model used in a recent $2 billion quantum computing award.

      SandboxAQ's CEO, Jack Hidary, refrained from disclosing the exact size of the government's stake, noting only that it does not come with voting rights or a seat on the board. The company counts among its investors prominent figures such as Google, Eric Schmidt, and Ray Dalio.

      Rather than offering a chatbot, SandboxAQ focuses on its "Large Quantitative Models," which are trained in physics and chemistry instead of human language. These models aim to evaluate millions of potential materials through software prior to any physical experimentation. The company asserts that this accelerates the discovery process from years to weeks, claiming its catalyst models, developed through 13.5 million calculations with Nvidia, operate approximately 20,000 times faster than traditional techniques, although these statistics are from SandboxAQ itself.

      However, this initiative is primarily a research investment, not a guarantee of a finished product. There are currently no new magnets or PFAS-free chemicals available, and the funding is directed at research rather than assured outcomes.

      An ironic point to consider is that the same administration funding the PFAS alternatives also eased some PFAS drinking-water deadlines last year. Nonetheless, the strategic reasoning is evident, especially as China oversees over 90 percent of the rare-earth magnets necessary for the production of America's most advanced chips, emphasizing the importance of supply-chain sovereignty.