To make genuine scientific breakthroughs, an AI needs to learn the abstract reasoning strategies and mental models of expert scientists. This involves teaching it higher-level concepts, such as thinking in terms of symmetries, a core principle in physics that current models lack.
Current LLMs fail at science because they lack the ability to iterate. True scientific inquiry is a loop: form a hypothesis, conduct an experiment, analyze the result (even if incorrect), and refine. AI needs this same iterative capability with the real world to make genuine discoveries.
Instead of relying on digital proxies like code graders, Periodic Labs uses real-world lab experiments as the ultimate reward function. Nature itself becomes the reinforcement learning environment, ensuring the AI is optimized against physical reality, not flawed simulations.
The ambitious goal of discovering a high-temperature superconductor isn't just a scientific target; it's a strategic choice. Achieving it requires building numerous sub-systems like autonomous synthesis and characterization, effectively forcing the creation of a general-purpose AI for science platform.
While pursuing a long-term research goal, the company's commercial strategy is to build AI co-pilots and intelligence layers for R&D workflows in established industries like space and defense. This approach productizes intermediate progress and targets massive existing R&D budgets.
In a field as complex as AI for science, even top experts know only a fraction of what's needed. Periodic Labs prioritizes intense curiosity and mission alignment over advanced degrees, recognizing that everyone, regardless of background, faces a steep learning curve to grasp the full picture.
Foundation models can't be trained for physics using existing literature because the data is too noisy and lacks published negative results. A physical lab is needed to generate clean data and capture the learning signal from failed experiments, which is a core thesis for Periodic Labs.
Simply scaling models on internet data won't solve specialized problems like curing cancer or discovering materials. While scaling laws hold for in-domain tasks, the model must be optimized against the specific data distribution it needs to learn from—which for science, requires generating new experimental data.