US project combines quantum AI and robotics for competitive rare earth supply chain

USA Rare Earth has partnered with French quantum computing company Pasqal and industrial AI specialist Riven Systems to develop more efficient ways of separating rare earth elements in the United States.

Announced on September 17, the project will use quantum machine learning and automated laboratory experiments to identify molecules that bind more effectively to rare earths.

The companies hope the approach will enable smaller processing facilities that consume less energy and cost less to operate.

Targeting a critical processing gap

Rare earth separation involves converting mixed material into individual oxides used across advanced manufacturing and other strategic industries. China currently leads this part of the supply chain, particularly for heavy rare earths such as dysprosium, terbium and yttrium.

“The key challenge the rare earth industry outside Asia faces is to separate the Mixed Rare Earth Carbonate (MREC) produced in upstream operations into individual, separated oxides,” said Alex Moyes, senior vice president of upstream operations in the US at USA Rare Earth.

Instead of relying primarily on lengthy trial-and-error testing, the partners intend to create a data-driven system for discovering extractants.

These chemical molecules selectively bind with particular rare earth elements during processing, helping operators separate them from mixed feedstocks.

A more effective extractant could reduce the number of processing stages required, as well as the amount of equipment and raw materials used. USA Rare Earth said this could lower both the capital and operating costs of its future facilities while reducing their environmental footprint.

Thousands of automated experiments

Under the planned project, Riven Systems will use its self-driving minerals-separation laboratory to conduct thousands of automated experiments.

The resulting chemical data will be used to train machine learning models that predict how selectively different extractants bind to rare earth elements.

Pasqal will then use its neutral-atom quantum processingunit to benchmark quantum machine learning models against models running on classical computers. The findings will help USA Rare Earth identify promising extractants for further testing.

“Rare earth materials are essential, and improving how they are processed has implications far beyond a single industry,” Pasqal CEO Wasiq Bokhari said.

The experiments will be tailored to materials that USA Rare Earth expects to process.

These include feedstock from the Round Top deposit in Sierra Blanca, Texas, third-party mixed rare earth carbonates and recycled cuttings, known as swarf, produced during magnet manufacturing.

Building an end-to-end discovery system

The initial machine learning project could eventually develop into an integrated extractant discovery pipeline spanning computer modeling, automated experiments, and physical validation.

Under that proposed system, Pasqal’squantum machine learning model would identify high-potential molecules before Riven tested them in its autonomous laboratory.

The strongest candidates would then be validated at USA Rare Earth’s research and development facility in Wheat Ridge, Colorado, before potentially being incorporated into the company’s processing flowsheets.

“AI and autonomous labs are the next frontier in critical mineral processing,” said Dr. Orion Archer Cohen, CTO and co-founder of Riven Systems.

The collaboration combines American rare earth and automation expertise with French quantum computing technology. Its broader objective is to help establish a more efficient and competitive Western supply chain for rare earth materials.