Materials for Extreme Environments.
Dynamical develops and supplies materials for power electronics, electric motors and turbines that run hotter and harder.
Research
Our research thesis, scientific meta-learning, is that agents can learn to choose the experiment that teaches the most and carry that skill to the next material.
Teaching AI to Explore the Physical World
SDL-2 measures whether AI agents can decide, from experiments they choose, whether a magnet process will reliably meet a stated performance requirement. Frontier models show an emerging ability to establish what an unfamiliar physical system can reliably do. The findings implicate reasoning under variation as a central capability for scientific exploration.
Scientific Autoresearch
The open-source Dynamical CLI is a unified interface for simulation and physical execution. In water electrolysis, agents selected the best catalyst 1.5× as often with Dynamical, while broader investigations traced how evidence shaped their scientific decisions.
Training Scientific Judgment from Physical Experiments
Post-training Qwen3.6-35B-A3B on recorded materials experiments, combined with evidence guidance and test-time scaling, produced 23% lower average log loss than the base system. Its judgments of experimental outcomes and confidence better matched the recorded results.
How Scientific Agents Learn from Experiments
SDL-1 turns recorded materials experiments into controlled tests of scientific learning. Across six agent systems, understanding an inspected experiment, improving predictions about other experiments, and choosing useful experiments were distinct capabilities.
Simulator-Verified Skill Acquisition for Scientific Instruments
Proprio turns instrument corrections into skills a new agent can pick up. New GPT-5.6 Luna agents given repaired skills succeeded in 6 of 6 simulated sessions against 4 of 6 with the originals. On a robot workstation, a new GPT-6 Astra agent given skills placed a charger in 2 minutes rather than 14.
Can a Self-Driving-Lab Agent Tell When the Evidence Is Enough?
VOE-Bench turns NIST manufacturing records into 104 evidence-review tasks. Across six models and 1,872 runs, 90% of final decisions met the review rules and were supported by inspected evidence, separating evidence acquisition from judgment.
Scaling Test-Time Verification for Novel Materials
We studied how property estimates can guide crystal generation with Crystalite and MatterGen. With Crystalite’s generator fixed, a trained probe raised predicted band-gap targeting from 0% to 24%; a separately trained checkpoint reached 43%.