Can AI design around scarce elements?

Introduction

The promise of AI-designed materials is not only that they could make technologies perform better. In some cases, the bigger question is whether they can make crucial technologies less dependent on fragile supply chains.

MatterGen, Microsoft’s generative materials-design system, offers a useful example. One of its showcased goals was not simply finding stronger magnets, but generating magnetic materials that combine useful performance with lower supply-chain risk. That matters because modern clean-energy systems, electric vehicles, robotics, advanced manufacturing equipment, data centres and defence technologies all rely heavily on permanent magnets whose production depends on a small number of countries and difficult-to-substitute elements.[Nature]nature.comA generative model for inorganic materials designby C Zeni · 2025 · Cited by 679 — Although MatterGen can be fine-tuned for any com…[IEA]iea.orgNew projects, partnerships and policies are needed to…Demand for magnet rare earths – notably neodymium, praseodymium, dysprosium a…

For advocates of an AI-enabled age of abundance, this is an important test case. The challenge is not merely inventing something better in a laboratory. It is whether AI can help civilisation escape bottlenecks created by scarce materials, geopolitical concentration, environmental costs and industrial vulnerability. MatterGen suggests one possible route: designing around the constraint itself.

Why material abundance is a supply-chain problem

Many visions of technological abundance assume that once humanity knows how to build something useful, scaling it becomes mostly an engineering challenge. In practice, physical supply chains often become the limiting factor.

Permanent magnets illustrate the problem. High-performance magnets used in electric motors, wind turbines, industrial machinery and many electronics frequently depend on rare earth elements such as neodymium, praseodymium, dysprosium and terbium. Demand for these magnet materials has risen rapidly alongside electrification and clean-energy deployment, with the International Energy Agency reporting that demand for key magnet rare earths has roughly doubled since 2015 and is expected to continue growing strongly.[IEA]iea.orgWith new export controls on critical minerals, supply…Oct 23, 2025 — The new rare earth export controls pose major risks for a range o…

The issue is not simply geology. Rare earth elements exist in many places, but mining, refining and magnet production are concentrated in a small number of locations. China dominates much of the processing and magnet-manufacturing chain, creating concerns about industrial resilience and geopolitical leverage. Multiple analyses estimate that Chinese firms control the majority of global rare-earth processing and magnet production capacity.[IDTechEx]idtechex.comGlobal rare earth mining, separation, refining, and magnet production supplyRare Earth Magnets 2026-2036: Technologies, Supply…This report characterizes rare earth magnet markets, technologies, and pla… Nature Recent export restrictions have turned what was once a largely technical discussion into a strategic one. Industry and policy reports in 2025[nature.com]nature.comA generative model for inorganic materials designby C Zeni · 2025 · Cited by 679 — Although MatterGen can be fine-tuned for any com… and 2026 describe disruptions, licensing delays and growing concern among manufacturers dependent on rare-earth magnets.[S&P Global]spglobal.com012726 rare earth supply bottlenecks set to persist in 2026S&P GlobalRare earth supply bottlenecks set to persist in 202627 Jan 2026 — China's rare-earth export restrictions are set to drive suppl…

This creates a broader question for AI-driven scientific acceleration. If advanced AI helps humanity discover better batteries, motors or energy systems, but all of them depend on vulnerable inputs, then material bottlenecks could slow the wider bloom of abundance. Scientific progress and supply-chain resilience become linked problems rather than separate ones.

How MatterGen combines performance and lower risk

Most materials-discovery systems are asked to optimise performance. MatterGen demonstrates a more ambitious idea: optimising performance and supply-chain characteristics simultaneously.

According to the Nature paper describing the system, researchers fine-tuned MatterGen to generate materials with both strong magnetic properties and lower supply-chain risk. The paper specifically highlights the challenge that many high-performance permanent magnets depend on rare earth elements whose supply chains are considered strategically vulnerable.[Nature]nature.comA generative model for inorganic materials designby C Zeni · 2025 · Cited by 679 — MatterGen is also able to design materials given multi…

This matters because traditional materials research often treats supply-chain concerns as a later-stage problem. A material might first be discovered because it performs well, only for industry to discover years later that it requires expensive, politically sensitive or environmentally difficult ingredients.

Inverse design changes the order of operations.

Instead of asking:

Can we find a high-performance magnetic material?

Researchers can ask:

Can we find a magnetic material that performs well while avoiding high-risk elements?

That sounds simple, but it greatly expands the search challenge. Materials scientists are no longer looking for a single optimum. They are balancing several competing objectives:

MatterGen’s significance is that it treats those tradeoffs as design inputs rather than after-the-fact constraints.[Nature]nature.comA generative model for inorganic materials designby C Zeni · 2025 · Cited by 679 — Although MatterGen can be fine-tuned for any com…

In principle, a future generation of materials models could search vast regions of chemical space for solutions that human researchers might never manually investigate. The objective would not necessarily be finding the strongest possible magnet. It would be finding the best overall compromise between performance, resilience and scale.

Why rare-earth-free magnets are so difficult

The challenge is that rare earth magnets became dominant for good reasons.

Neodymium-iron-boron magnets and related technologies achieve extremely high magnetic performance. They are compact, powerful and well suited to electric motors, wind turbines and many advanced applications. Replacing them is not as straightforward as swapping one ingredient for another.[ScienceDirect]sciencedirect.comRare earth permanent magnets for the green energy…by Y Ghorbani · 2025 · Cited by 111 — This review paper provides an ove…

Researchers have spent years exploring alternatives, including ferrites, alnico materials, manganese-based compounds, iron-nitride systems and Heusler alloys. Many of these approaches reduce reliance on critical rare earths, but often at the cost of weaker magnetic performance, more difficult manufacturing requirements or lower durability under real-world conditions.[arXiv]arxiv.orgarXiv Matter Gen: a generative model for inorganic materials designMatterGen: a generative model for inorganic materials designby C Zeni · 2023 · Cited by 230 — We present MatterGen, a model that generate…

This is where AI-assisted design could matter most.

A human researcher might examine thousands of candidate compounds. A generative model can explore vastly larger regions of possibility space and propose unconventional combinations that satisfy multiple constraints simultaneously. That does not guarantee success, but it increases the chance of finding useful compromises.

The broader significance for an AI-bloom worldview is that abundance is frequently blocked by optimisation problems rather than fundamental physical impossibilities. The world may not need a perfect substitute for every scarce material. It may need substitutes that are good enough while being easier to source, manufacture and scale.

The hidden tradeoff: scarcity can move rather than disappear

There is a temptation to assume that removing a risky element automatically solves the supply problem. Reality is usually more complicated.

A magnet designed without dysprosium or terbium may still depend on other materials that have their own supply constraints. It may require more material overall, more energy-intensive processing or more complex manufacturing techniques. A substitute could reduce geopolitical dependence while increasing environmental costs or production expenses.

This is a recurring pattern in energy and materials transitions.

For example:

Researchers studying rare-earth supply systems increasingly emphasise that risk exists at multiple levels: mining, refining, manufacturing, transport and geopolitical relationships. Simply changing the chemistry does not automatically remove all vulnerabilities.[IDEAS]ideas.repec.orgIDEAS/RePEcGeopolitical risk and the global supply of rare earth permanby L Depraiter · 2025 · Cited by 47 — This paper delves deep into…[RePEc]ideas.repec.orgIDEAS/RePEcGeopolitical risk and the global supply of rare earth permanby L Depraiter · 2025 · Cited by 47 — This paper delves deep into…

This is one reason MatterGen’s multi-objective approach is notable. The goal is not merely replacing one element. It is searching for solutions that perform well across several dimensions of risk simultaneously.

What industry would still need before adoption

Even if AI generates promising magnetic materials, several difficult stages remain.

Laboratory validation

Generated materials must first be synthesised and tested. A material that appears stable in simulations may prove difficult to manufacture, expensive to process or unreliable under industrial conditions.

The history of materials science is full of theoretically attractive compounds that never became commercial products because practical realities intervened.

Manufacturing compatibility

Factories represent huge capital investments. A new magnet material must fit into existing production systems or justify the cost of replacing them.

A material that performs slightly better on paper may fail commercially if manufacturers need entirely new equipment, supply contracts or quality-control processes.

Long-term reliability

Magnets used in vehicles, power systems and industrial equipment must operate for years under heat, vibration and mechanical stress.

That means researchers need evidence not only of magnetic performance but also of durability, corrosion resistance and predictable behaviour across long operating periods.

Economic competitiveness

Even strategically valuable materials face cost pressures.

Manufacturers may tolerate higher costs for defence applications or critical infrastructure, but mass-market adoption usually requires attractive economics alongside technical performance.

This creates an important reality check for AI-driven discovery. Generating a candidate material is not the same as transforming an industry. Discovery is only the first step in a chain that includes testing, manufacturing, certification, financing and deployment.

A glimpse of a broader abundance strategy

The magnet example is interesting because it reveals a deeper possibility behind inverse materials design.

Many future technologies may be limited less by ideas than by physical bottlenecks. Clean energy systems require minerals. Advanced computing requires specialised materials. Robotics depends on motors, sensors and industrial inputs. Even ambitious visions of space industry eventually run into questions about matter, manufacturing and supply chains.

MatterGen hints at a future where AI helps search not only for higher performance, but for resilience itself.

If such systems mature, researchers could increasingly specify goals such as:

That would not eliminate scarcity. Physics, economics and geopolitics would still matter. But it could shift some constraints from fixed limits into engineering problems.

For the larger AI-bloom vision, that distinction is significant. A flourishing future depends not only on creating powerful technologies, but on making them deployable at civilisational scale. The ability to design around bottlenecks may ultimately matter almost as much as the ability to invent new capabilities in the first place. The magnet case is a small but concrete example of how AI-assisted discovery could begin to move from chasing peak performance towards something closer to sustainable abundance.

Amazon book picks

Further Reading

Books and field guides related to Can AI design around scarce elements?. Use these as the next step if you want deeper reading beyond the article.

Book

Human Compatible

By Stuart Russell

Explores how powerful AI could advance humanity if aligned with human values.

See on Amazon

Book

The Coming Wave

By Mustafa Suleyman

Directly addresses whether advanced AI can deliver broad civilisational benefits while remaining governable.

See on Amazon

Book

The Premonition

By Michael Lewis

Explores early warning, forecasting and public-health response.

See on Amazon

Book

Apollo's Arrow

By Nicholas A. Christakis

Examines societal responses to pandemics and preparedness.

See on Amazon

Browse more on Amazon:Human CompatibleThe Coming WaveThe Premonition

As an Amazon Associate I earn from qualifying purchases.

eBay marketplace picks

Marketplace Samples

Topic-anchored marketplace searches for visual, collectible, or second-hand items related to this page.

Shop location

UsingUSA USA USAUKCanadaAustraliaIreland

Example eBay listing

20mm x 3mm Magnets Neodymium Disc 20 x 3 Rare Earth Magnet Strong N52 Grade

SearcheBay.co.uk: rare earth magnet

Browse similar oneBay.co.uk

Example eBay listing

Neodymium Magnets Extra Strong Rare Earth Magnets For Fridge Whiteboard

SearcheBay.co.uk: rare earth magnet

Browse similar oneBay.co.uk

Example eBay listing

Small N52 strong neodymium fasteners disc 2mm 3mm 4mm 5mm 6mm rare earth magnets

SearcheBay.co.uk: rare earth magnet

Browse similar oneBay.co.uk

Example eBay listing

Rare Earth DIY Industrial Strong Ring Neo Neodymium Disc Round Magnet with Hole

SearcheBay.co.uk: rare earth magnet

Browse similar oneBay.co.uk

Browse more oneBay.co.uk

Example items shown for inspiration; availability and pricing can change. Branchoria may earn a commission if you purchase through outbound eBay links.

Endnotes

1. Source: nature.com
Link:https://www.nature.com/articles/s41586-025-08628-5

2. Source: iea.org
Link:https://www.iea.org/news/new-projects-partnerships-and-policies-are-needed-to-address-supply-chain-risks-for-rare-earth-elements

3. Source: sciencedirect.com
Link:https://www.sciencedirect.com/science/article/pii/S0921344924005573

4. Source: idtechex.com
Title: Global rare earth mining, separation, refining, and magnet production supply
Link:https://www.idtechex.com/en/research-report/rare-earth-magnets/1112

5. Source: nature.com
Link:https://www.nature.com/articles/s41586-025-08628-5_reference.pdf

6. Source: arxiv.org
Title: arXiv Matter Gen: a generative model for inorganic materials design
Link:https://arxiv.org/abs/2312.03687

7. Source: arxiv.org
Link:https://arxiv.org/abs/1909.03275

Published: September 7, 2019

8. Source: arxiv.org
Title: arXiv Permanent Magnets Based on Hard Ferrite Ceramics
Link:https://arxiv.org/abs/2310.02106

9. Source: ideas.repec.org
Link:https://ideas.repec.org/a/eee/eneeco/v146y2025ics0140988325003202.html

10. Source: sciencedirect.com
Link:https://www.sciencedirect.com/science/article/pii/S0140988325003202

11. Source: microsoft.com
Title: mattergen a new paradigm of materials design with generative ai
Link:https://www.microsoft.com/en-us/research/blog/mattergen-a-new-paradigm-of-materials-design-with-generative-ai/

12. Source: microsoft.com
Link:https://www.microsoft.com/en-us/research/video/mattergen-a-generative-model-for-materials-design/

13. Source: iea.org
Link:https://www.iea.org/commentaries/with-new-export-controls-on-critical-minerals-supply-concentration-risks-become-reality

14. Source: iea.org
Link:https://www.iea.org/reports/rare-earth-elements/executive-summary

Published: April 2025

15. Source: spglobal.com
Title: 012726 rare earth supply bottlenecks set to persist in 2026
Link:https://www.spglobal.com/energy/en/news-research/latest-news/metals/012726-rare-earth-supply-bottlenecks-set-to-persist-in-2026

Additional References

16. Source: sheffield.ac.uk
Link:https://sheffield.ac.uk/royce-institute/impact-and-outreach/impact/case-studies/collaboration-rapidly-develops-rare-earth-free-materials

17. Source: iuk-business-connect.org.uk
Link:https://iuk-business-connect.org.uk/wp-content/uploads/2022/06/Final_Magnets-Document_Innovate-UK-KTN.pdf

18. Source: facebook.com
Link:https://www.facebook.com/microsoftresearch/posts/microsoft-researchers-introduce-mattergen-a-model-that-can-discover-new-material/1040950821397547/

19. Source: sites.lsa.umich.edu
Link:https://sites.lsa.umich.edu/mje/2026/01/09/market-concentration-of-rare-earth-elements-chinas-dominance-and-the-global-response/

20. Source: globalpolicywatch.com
Link:https://www.globalpolicywatch.com/2026/02/heavy-rare-earth-elements-rising-supply-chain-risks-and-emerging-policy-responses/

21. Source: csis.org
Title: chinas new rare earth and magnet restrictions threaten us defense supply chains
Link:https://www.csis.org/analysis/chinas-new-rare-earth-and-magnet-restrictions-threaten-us-defense-supply-chains

22. Source: sfa-oxford.com
Title: sfa china s rare earth export controls and their impact on global supply chains
Link:https://www.sfa-oxford.com/market-news-and-insights/sfa-china-s-rare-earth-export-controls-and-their-impact-on-global-supply-chains/

23. Source: advisorpedia.com
Link:https://www.advisorpedia.com/viewpoints/microsofts-mattergen-a-game-changer-in-material-discovery/

24. Source: youtube.com
Link:https://www.youtube.com/watch?v=Smz1go6_Spo

25. Source: md-one.com
Title: materials nexus discovers new rare earth free magnet using ai algorithm
Link:https://www.md-one.com/news/materials-nexus-discovers-new-rare-earth-free-magnet-using-ai-algorithm