Hello readers,
Welcome to the AI For All newsletter! Today, we’re talking about AI’s attempts to render rare earth metals moot, connectivity for physical AI, and more!
AI in Action: AI is racing to replace the world's most contested metals
For two decades, finding a permanent magnet that skips rare earth elements has meant a slow grind: synthesize a material, test it, learn a little, try again. Ames National Laboratory scientist Prashant Singh is proposing a different route. His team is pairing physics-based modeling and high-throughput simulation with reasoning AI tools that can flag promising candidates before anyone sets foot in a lab. The work feeds into the Department of Energy's Genesis Mission, which is trying to secure the country's supply of rare earths and other critical minerals.
The magnets that power everything from wind turbines to missile guidance systems get their strength from rare earth elements, which are costly and largely sourced overseas. Singh's approach models how a material's atomic structure and electron behavior shape magnetic strength, energy storage, resistance to demagnetization, and heat tolerance — the physics that decides whether two elements will actually make a good magnet together. Instead of guessing and testing, researchers can narrow the field computationally first.
What makes the models useful, Singh says, is training them on real experimental and calculated material data rather than generic datasets, so predictions stay grounded instead of drifting outside what's actually known. The tools also weigh something a lab bench can't: supply and cost. Singh's team is building on an earlier tool called DuctGPT, originally developed to find materials for fusion reactors, to let researchers pose design questions and get answers that account for which materials are affordable and available to source domestically, not just which ones look good on paper.
🔥 Rapid Fire
Commentary: The Subprime Data Center Crisis
“The nature of these SPVs makes it difficult to quantify the exact scale of data center debt, but Bloomberg estimates that there’s over $500 billion in outstanding AI data center debt, with (per Garima Kapoor of Elara Securities Research) at least $200 billion of it held by private credit, making up roughly 8% of outstanding private credit loans.”
“That being said, the number is likely much higher. Nikkei Asia reported this week that Meta, Google, Amazon, Microsoft, and Oracle have accrued around $1.65 trillion in outstanding debt in the last five years, with an additional hundreds of billions of dollars’ worth of off balance sheet debt, meaning that the corporate structure allows the company to not include it as part of its liabilities.”
Oracle could face $7 billion collateral bill for Wisconsin data center
“The Public Service Commission of Wisconsin, which is responsible for scrutinizing and setting the prices the state’s utilities can charge, has declined to reconsider rules it imposed on utility We Energies, which would require Oracle to provide a $7 billion letter of security at a yearly cost of more than $100 million.”
What awaits Larry Ellison: 1.2 million South Korean investors margin called
Meta is planning to sell excess compute to Anthropic
This is further confirmation that there is little demand for AI compute outside OpenAI and Anthropic
Alphabet falls as $205 billion spending plan fuels AI cost fear
Stocks hit by AI doubts — SpaceX hits all-time low
The Army is burning through its AI tokens, told to limit use
Startups urge government not to ban Chinese open models
Commenatry: The AI Productivity Illusion
Commentary: AI Mania Is Eviscerating Global Decision-Making
Google's AI search functions pose ‘unacceptable risk’ to children
How to find out what ChatGPT and Gemini know about you
📖 What We’re Reading
This is no longer just about copilots, chatbots, or even humanoid robotics. It is about something much bigger: AI moving into the physical world at scale. From excavators and surgical systems to robotaxis, autonomous forklifts, rail inspection systems, and precision agriculture equipment, intelligence is now being embedded directly into the machines that move industries forward.
