Highlights
- Rahul Mewawalla highlights rare earth elements as strategic assets with significant geopolitical implications for technological dominance.
- China controls over 90% of rare earth magnet manufacturing, creating potential supply chain vulnerabilities for AI and tech infrastructure.
- Real risk lies in magnet-to-device midstream dependencies, with potential significant impacts on tech sectors by 2026 if export controls tighten.
Rahul Mewawalla’s op-ed via Broadband and Breakfast (opens in a new tab) on rare earth minerals and their impact on AI and digital infrastructure offers a timely, high-level warning—but it also conflates key materials, misidentifies supply chain bottlenecks, and underplays the structural timeline of risk. Here’s what he gets right, and more importantly, what he gets wrong.
On the Money
Rahul Mewawalla is CEO and President of NASDAQ-listed Mawson Infrastructure Group, and formerly an executive with technology companies such as Yahoo, Nokia, and General Electric Company, cites the obvious, that rare earth elements (REE) represent a major national security issue.
Mewawalla correctly recognizes that REEs have become strategic assets in the geopolitical race for technological dominance. This is especially true for dysprosium, terbium, and neodymium, which are essential for permanent magnets in motors, generators, and some laser systems. Their control is tightly concentrated in China.
Other points to consider include vulnerabilities associated with AI/Data center supply chain. He rightly notes that while 2025 is unlikely to see acute REE-related disruptions, by 2026-2027, supply constraints, cost increases, or political restrictions could begin filtering into the hardware stack, especially as demand surges for electric cooling fans, servo motors, and even drone-grade sensors.
Finally, Mr. Mewawalla points out that existing multi-month and multi-year procurement buffers in the tech sector mask deeper vulnerabilities, giving a false sense of security. That’s accurate, and important, and we applaud him for calling them out.
What’s of Concern?
While Mewawalla names REEs like gadolinium, lutetium, samarium, and scandium, these are not primary inputs in mainstream AI, cloud, or data center hardware. The real bottlenecks are:
- Neodymium (Nd) and Praseodymium (Pr) – for permanent magnets in cooling fans, power conversion, and robotics
- Dysprosium (Dy) and Terbium (Tb) – forhigh-temperature magnets
- Yttrium (Y) – sometimes in phosphors and ceramics
Misidentifying niche rare earths like lutetium or gadolinium, while ignoring the dominant NdFeB magnet market, shows a lack of depth in the technical dependencies. AI systems are not failing because of a scandium shortage.
What’s the real risk? How about magnet-to-device midstream?
The real choke point isn’t the mine—it’s the magnet supply chain, especially sintered NdFeB and bonded magnets, nearly all of which are produced in China. Mewawalla’s piece treats rare earths as raw inputs without understanding that China controls over 90% of rare earth magnet manufacturing. This midstream dependency is where the U.S. and EU are most exposed, not just to cost hikes, but outright embargoes.
The expert does not mention China’s export controls. Perhaps the most glaring omission is the failure to cite China’s 2023-2025 export restrictions on gallium, germanium, and rare earth permanent magnet alloys. These have already impacted Japanese and European manufacturers, and are likely to worsen as the U.S. imposes reciprocal tariffs under IEEPA.
Then there is the glossing over of timelines for new supply. Mewawalla urges companies to “find new opportunities today,” but doesn’t acknowledge that building new REE mining, separation, and magnet facilities takes 5–10 years, not 12 months. The notion that companies can “assess alternative suppliers” by 2026 and meaningfully scale by 2027 is wishful thinking without federal industrial coordination.
Hype vs.Reality
Mewawalla is correct in identifying rare earths as a brewing constraint on future tech infrastructure. But his analysis lacks specificity, misrepresents the real REE-to-hardware connection, and fails to articulate the true geopolitical and industrial stakes. It reads more like a cautious investor letter than a serious technical or strategic briefing.
Bottom Line for Investors
- Yes, rare earth access will shape the future of AI hardware and national power.
- No, you won’t see the impact in Q3 earnings calls—but you will by 2026 if China tightens magnet exports or if EV, defense, and tech sectors converge on the same constrained inputs.
- Watch for U.S.-EU-Japan coordination on magnet manufacturing and any extensions of the DPA/CHIPS Act that subsidize REE-to-magnet supply chains.
Until then, invest with precision and be cautious about taking things at face value. All of us are learning in defense of Mr. Mewawalla. He has some very good points. Note, some of the top experts we know in the rare earth element space inform us they are learning something new every day, three decades later!
Rare Earth Exchanges (REEx) will continue to track the facts and separate signal from speculation. Check out our Forum (opens in a new tab) for the chatter.
Leave a Reply