SandboxAQ and Jack Hidary on rare earths

Can AI and quantum computing reduce supply chain risk?

SandboxAQ and Jack Hidary on rare earths: can AI and quantum computing reduce supply chain risk?

By SandboxAQ editorial team | Last updated:

A South China Morning Post explainer asks whether the US can loosen concentration risk in rare earth minerals by using AI and quantum computing to develop synthetic substitutes or alloys faster than new mines can come online. In the piece, Jack Hidary, CEO of SandboxAQ, says these technologies could cut the time required to secure critical materials to “just a few years” — potentially bypassing the traditional 10 to 20 years needed to bring a new mine online.

What SCMP reported: the bottleneck is processing, not just mining

SCMP frames the challenge as structural: China controls most of the world’s rare earth mining and nearly 90% of processing and refining capacity, which has made independent supply chain buildout difficult for Western governments. The article also notes why processing is hard to replicate quickly — China captured the processing market partly by absorbing the environmental and health risks associated with chemically hazardous refining that many Western nations avoided.

Jack Hidary’s core claim: substitutes and alloys can move faster than mines

Hidary’s argument, as reported by SCMP, is that “the route to challenging China’s supremacy” is not solely about opening new mines. Instead, it’s using AI and quantum tools to synthesize substitutes or alloys — a pathway that could compress timelines from decades to years.

The important nuance: this is not “AI replaces mining.” It is advanced computation accelerating engineering around constraints, which can reduce dependence on the most bottlenecked parts of the supply chain before new industrial capacity is ever built.

Where AI helps first in materials discovery

AI tends to help most when the problem space is large and experimentation is expensive. In materials discovery, that means screening large design spaces for candidate compositions, narrowing down which options are worth real-world testing, and modeling tradeoffs between performance, cost, and manufacturability. SandboxAQ’s Large Quantitative Models are designed for exactly this kind of physics-grounded scientific problem. More on the approach in the context of AI for science and discovery.

In supply chain terms, the value is speed: a faster path to viable alternatives can relieve pressure long before new industrial capacity comes online.

Where quantum computing fits

SCMP positions quantum computing as part of the toolset alongside AI. In practice, quantum is discussed as potentially useful for certain simulation and optimization workloads — part of an advanced computation stack aimed at reducing trial-and-error cycles. The practical takeaway is measured: if quantum methods expand the set of problems that can be modeled efficiently, they could contribute to faster iteration loops in some materials workflows.

The counterpoint: lab breakthroughs don’t automatically become industrial capacity

SCMP includes a clear caveat from analysts: even if AI-accelerated chemistry delivers promising substitutes, converting laboratory breakthroughs into large-scale industrial processing and manufacturing is still difficult. China’s advantages were “perfected over decades,” and technical advances alone may not offset those structural advantages quickly.

That’s the central tension: discovery can speed up, but industrialization and adoption still determine whether supply chain risk meaningfully declines.

What to watch over the next 6–12 months

Credible progress on substitutes: candidate materials or alloys that meet real performance requirements, not just lab results
Scale-up evidence: whether candidates move beyond lab settings into manufacturable processes
Processing investment: parallel buildout of non-China refining and processing capacity
Downstream adoption: whether manufacturers begin referencing alternatives in actual procurement or design decisions

FAQ

What are rare earth minerals and why do they matter?

Rare earths are critical inputs for many advanced technologies. The supply chain risk comes from concentrated mining and especially concentrated processing and refining capacity.

Why is processing and refining the chokepoint?

SCMP reports that China controls nearly 90% of processing and refining, and that processing carries environmental and health burdens that have historically discouraged replication elsewhere.

Can AI really reduce rare earth dependence?

AI can help accelerate materials discovery by screening candidates and narrowing experiments, which may speed up the development of substitutes and alloys that reduce reliance on constrained inputs.

Will quantum computing help AI in materials science?

It’s discussed as part of an advanced computation toolkit that could assist certain simulation and optimization workloads, potentially shortening iteration cycles in some cases.

How long could this take compared with opening new mines?

SCMP reports Hidary’s view that substitutes and alloys could take “just a few years,” versus 10 to 20 years for a new mine to come online. The article also emphasizes the real challenge of scaling lab advances into industrial manufacturing.

Related resources from SandboxAQ:

‍