Bezos-backed CuspAI raises $2.6B to use AI for chip materials

CuspAI, a startup using artificial intelligence to discover new materials for computer chips, has raised $450 million in funding at a $2.6 billion valuation. The round was led by a handful of private investors including Bezos Expeditions, the venture fund of Amazon founder Jeff Bezos. Chipmakers Nvidia and Advanced Micro Devices (AMD), along with Meta Platforms, are serving as developmental partners with the company.

The company's core mission addresses a growing constraint in the AI hardware supply chain. As AI computing infrastructure expands, demand for critical materials is straining global supplies. The International Energy Agency predicts that data centers' demand for gallium in 2030 could be 10% greater than today's supply. More pressingly, China controls 99% of the planet's refined gallium supply, creating geopolitical vulnerability for the rest of the world. Beyond gallium, other elements essential to AI computing platforms—iridium and tantalum—are also becoming increasingly difficult and costly to obtain, limiting the industry's ability to scale.

CuspAI's AI platform addresses this bottleneck by identifying alternative materials that could substitute for scarce elements and developing new syntheses for materials that have previously been impossible to produce. The company cannot create or procure the metals themselves, but it can virtually model and test alternatives at a fraction of traditional cost and timeline. Work that would normally take years and millions of dollars to complete can now be done in months at a fraction of the cost. CuspAI is rooted in a real supply-chain need rather than speculative technology; industry forecasts from both Mordor Intelligence and Precedence Research predict that the generative AI material science industry will grow at an average annual pace of more than 24% at least through 2030.

Bezos' involvement in CuspAI is not his first investment in AI-enabled manufacturing. In November, Bezos Expeditions invested in Prometheus, an AI startup that helps engineers design and manufacture complex physical products, potentially including microchips made from materials discovered by CuspAI. The two investments appear complementary, signaling confidence that AI-driven material science will become a durable and valuable capability. CuspAI is not publicly traded, and there are no whispers of an IPO in the foreseeable future. Still, if the company succeeds in bringing cost-effective alternatives to market within five to ten years, it could prove transformative for chipmakers like Nvidia, AMD, and others that depend on expensive or geopolitically vulnerable materials.