Winning the AI Race Part 3: Jensen Huang, Lisa Su, James Litinsky, Chase Lochmiller

The first conversation in this part opens by introducing James (Jim) Litinsky as the founder and CEO of MP Materials. The introduction says he was running a hedge fund that had invested in Molycorp, a company that went out of business. Rather than walk away, he shut down the fund and took over the business. The interviewer says MP Materials is now the largest, and possibly the only, U.S. company that supplies and refines rare earth materials and makes magnets. The guest's answer is that MP represents "100% of the American industry." The introduction also points to two recent deals: a public-private partnership with the Department of Defense worth around $400 million, and a large agreement with Apple.

So rare earth magnets are really the feedstock to physical AI.

The guest's case for why rare earths matter is that electrified motion depends on rare earth magnets. That includes robots, drones and the other products grouped under physical AI, which the guest calls the biggest industry still to come. The site is at Mountain Pass, California, about a 45-minute drive from the Las Vegas Strip and visible from the road. The guest describes it as the best rare earth ore body in the world. The takeover is dated to 2015, with the assets bought out of bankruptcy in 2017 after what is described as a two-year battle.

The main technical point is that mining is only the first step. Rare earths have to be separated in a specialty chemical process, which the guest likens to a multi-billion-dollar refinery. The separated material then has to be turned into metal and then into magnets. The guest argues that every layer matters. A country with all the rare earths but no magnet production ends up "sending it to China." A country with magnet capability but no rare earths ends up "relying on China."

The guest rejects the view, held when the assets were bought, that a U.S. operation could not compete with China. The account is that the site was world-class but the process flow had to be reorganized and the company had to invest further downstream. The figures given are about $1 billion invested over eight years, a public listing in 2020, the build-out of refining capacity, and a magnet factory in Texas announced about four years ago. GM is the foundational customer. The factory is producing auto-grade magnets to GM's specification, and magnet sales to GM are expected to ramp up at the end of this year.

The guest describes the DoD partnership as having three pillars. First, DoD becomes the company's largest economic investor. Second, DoD provides a price floor for MP's commodities, so that what the guest calls Chinese mercantilism cannot push prices below the cost of production. Third, the investment speeds up the build-out of magnet manufacturing. The Texas facility is being expanded for Apple. A separate new facility is meant to multiply capacity by ten, with DoD as the 100% offtake customer and a 50-50 profit-sharing partner.

In the discussion, the arrangement is presented as an investment, not a handout. The government holds equity and warrants, gains if commodity prices rise, and buys all the output. The guest says the deal guarantees a level of profit to justify building the facility, and above a certain threshold the two sides split the economics 50-50. One participant notes that this effectively means the taxpayer is a participant.

This is private risk, public risk, public upside, private upside.

The guest contrasts the deal with past government efforts to back an industry and "pick a winner." In those cases, the guest says, money was handed over and the result was public risk with private upside. The guest predicts that in five years the deal will be seen as a first-of-its-kind model in which the taxpayer made money, and asks to be held to that. The stated reasoning is a form of "mutually assured economic destruction": if China believes the United States also has national champions, there is no longer any point in subsidizing sales to the rest of the world. Prices could then normalize, and U.S. firms would be freer to invest and expand. This is presented as a hope, not a result: "I hope I'm right on this."

Asked why the company went to the government instead of private markets, the guest names mercantilism as the reason, comparing the need for government support and credit to World War II and the railroad boom.

Because the Chinese will sell magnets for below the cost of raw materials

The guest explains that any competitor that makes progress can be put out of business overnight, which discourages investment. As fiduciaries to shareholders, the company says it could not commit to the scale and timeline DoD wanted without assurance that it would not be destroyed by mercantilist pricing and that it would have a buyer for the magnets.

The guest says the negotiation was not easy. The deal was led by DoD under what the guest calls a mandate directly from the president, and the guest praises the Pentagon's leadership. The guest calls it the hardest work of their life, an aggressive negotiation in the style of private equity or distressed lending, and notes that the transaction documents are public. The guest's reading of DoD's intent, while saying they do not speak for DoD, is that the government removed the risks the company cannot control, such as mercantilism and certain customer issues. In return, the company is held to account for what it can control: executing on an aggressive timeline and keeping costs down. If the company gets its costs wrong, it bears that exposure.