USA Rare Earth's $1.2 billion magnet plant starts commissioning in 2028

The person this lands on is the engineer with one magnet supplier on the bill of materials and no approved alternate. The date matters to them more than the capital number. USA Rare Earth expects commissioning at Blacksburg to begin in 2028 ^[4], and commissioning is when a line starts running, not when it runs at 6,400 metric tons a year ^[3].

The federal package is larger than the plant. Commerce is backing the project with up to $1.6 billion in CHIPS Act support, $277 million of it in direct grants and $1.3 billion in senior secured loans ^[6], against a build the company prices at $1.2 billion ^[1]. Those two lines add to $1.577 billion ^[1], and the support ceiling is about 1.33 times the stated build cost ^[2]. Most of it is debt the company repays ^[6].

Divide the build cost by the magnet target and Blacksburg costs roughly $187,000 for each annual metric ton of magnet capacity ^[3]. The same site is also meant to make 5,000 metric tons a year of strip-cast metal and alloy ^[3], so that per-ton figure is buying two product lines, not one. It is also about $2.4 million of capital for each of the 490 jobs ^[5].

The 10,000 metric ton number the company uses for domestic NdFeB capacity is Blacksburg plus a planned expansion at Stillwater, Oklahoma ^[11]. Subtract Blacksburg's 6,400 and 3,600 metric tons sit with that expansion ^[4]. Stillwater's first commercial production line was commissioned earlier this year ^[10]; the press release does not give its current output or a ramp curve for Blacksburg.

Gregory Bowman, the company's chief global policy officer, said: "A secure rare earth supply chain isn't built with a single mine or a single factory. It requires rebuilding every link." Barbara Humpton, the chief executive, said of the groundbreaking that "it moves our vision from plans on paper to infrastructure taking shape." The company lists defense, aerospace, semiconductor manufacturing, physical AI, mobility, energy and healthcare among the industries that depend on these materials ^[13].

Two questions sort a hardware team's exposure. Does the program's first shipment land before or after 2028 ^[4]. And can the design absorb a different magnet grade and vendor without a requalification that costs more than the supply risk it removes. Programs shipping before 2028 with tight magnetic margins get nothing from Blacksburg, and the working answer for them is inventory or a second import source. Programs shipping after, on designs that tolerate substitution, are where naming USA Rare Earth in the qualification plan is cheap now and useful later. The teams to worry about are the ones treating a groundbreaking as a change in their supply position; that arrives when parts with a data sheet come off the line in 2028 ^[4].