America needs critical minerals and grid upgrades for AI infrastructure
Original: Energy, Minerals, and the Physical Stack Behind AI
Source: a16z.simplecast.com ↗
Who: Posted on the a16z podcast — a16z is a major Silicon Valley venture capital firm. The episode is hosted by Erin Price-Wright (a16z partner) and features Turner Caldwell (founder of Mariana Minerals) and Drew Baglino (long-time Tesla executive turned energy infrastructure founder).
What's new: The core argument is that hitting physical limits is not a software problem — it is a mining, refining, and electricity delivery problem. The US is more than 50 years behind China in building the supply chains for critical minerals (materials like lithium, cobalt, and rare earths used in batteries, chips, and motors), and the American power grid runs on equipment designed a century ago.
How it works: Caldwell explains that Mariana Minerals uses autonomous systems and to remove the expert know-how bottleneck from mineral processing — meaning a machine can learn the tricky refining steps that previously required scarce human specialists. Baglino argues that aging mechanical grid equipment, particularly large transformers (the heavy devices that step electricity up or down in voltage before it reaches homes and factories), should be replaced with — silicon and software doing the job of iron and oil.
Why it matters: Co-locating the full supply chain — mining, refining, manufacturing — inside the US compresses the timeline and reduces fragility more than chasing cheaper labor abroad. Both founders draw on lessons from Tesla: willingness to own the whole production stack rather than outsource critical steps, and recruiting people motivated by the mission rather than just the salary. Smarter federal permitting and a clear grid-investment framework, they argue, would unlock progress faster than any single technology.
Caveats: The episode is a founder-friendly venue with no adversarial pushback; timelines for permitting reform and grid modernization are asserted with optimism but without detailed evidence. The claim that can fully replace specialist know-how in mineral refining is plausible but unproven at commercial scale.