Radical AI autonomous alloy labs: implications for critical minerals supply
Reviewed by Tom Sullivan
First reported on MINING.com
30 Second Briefing
Radical AI has built a commercial “self‑driving lab” at Brooklyn Navy Yard that uses AI, robotics and an active learning loop to design, manufacture and test 50–200 g alloy samples, generating 300 new compositions in about 16 weeks for aerospace and defence clients. Backed by RTX and Nvidia and partnering with Sandia National Laboratories on metal–hydrogen interactions, the platform lets users exclude supply‑constrained elements such as hafnium while targeting specific strength, oxidation and hypersonic‑relevant high‑temperature properties. By collapsing 10–20 year alloy development cycles to roughly 1–3 years, the approach could reshape demand patterns for certain critical minerals.
Technical Brief
- Autonomous lab synthesises 50–200 g alloy buttons, then measures chemistry, grain structure and mechanical response.
- Test regime includes hardness, tensile strength and elevated-temperature performance for hypersonic-relevant conditions.
- Active learning loop retrains machine-learning models after each experimental batch, tightening composition–property predictions over time.
- Users can hard‑exclude specific elements (e.g. hafnium) while constraining strength, durability, oxidation and temperature limits.
- Radical AI reports high‑temperature alloys with minimal or near‑zero hafnium content for aerospace and defence environments.
- Sandia collaboration targets alloys for metal–hydrogen interactions and hydrogen separation from waste gas streams.
- Platform operates purely at discovery scale; customers retain IP and must separately qualify alloys for certification.
- For mining and critical minerals, accelerated substitution R&D could reweight long‑term demand for specific alloying elements.
Our Take
The use of autonomous labs for critical minerals like hafnium positions Radical AI at the more experimental end of the 48 Materials stories in our database, where most AI-related items still focus on mine planning and processing optimisation rather than front‑end alloy discovery.
Compressing materials development from ~20 years to a 1–3 year window, as targeted here, would materially change project sequencing for US-based metals and critical minerals developers, allowing design teams to iterate alloy specifications in parallel with early-stage mine or recycling flowsheet design instead of locking in decades-old standards.
Locating this capability at the Brooklyn Navy Yard in New York gives US defence and aerospace OEMs such as RTX a domestic testbed for small-batch (50–200 g) alloy prototyping, which could reduce reliance on China-centred supply chains for niche hafnium-bearing and other advanced metal systems.
Prepared by collating external sources, AI-assisted tools, and Geomechanics.io’s proprietary mining database, then reviewed for technical accuracy & edited by our geotechnical team.
Related Articles
Related Industries & Products
Mining
Geotechnical software solutions for mining operations including CMRR analysis, hydrogeological testing, and data management.
CMRR-io
Streamline coal mine roof stability assessments with our cloud-based CMRR software featuring automated calculations, multi-scenario analysis, and collaborative workflows.
HYDROGEO-io
Comprehensive hydrogeological testing platform for managing, analysing, and reporting on packer tests, lugeon values, and hydraulic conductivity assessments.
GEODB-io
Centralised geotechnical data management solution for storing, accessing, and analysing all your site investigation and material testing data.