Announcement
Shodh AI, an Indian AI‑Mission‑backed startup, announced the creation of LUCAN, a physical‑AI foundation model that can reason across scales from molecular behavior to manufacturing equipment.
Background
The company was founded by Dr. Arastu Sharma after research at the University of Cambridge and Microsoft Research and work with India’s DRDO defence‑AI ecosystem. It was selected under the Government of India’s IndiaAI Mission to develop foundation‑model capabilities for science, which funded the first generation of the model. The government is now providing additional support to scale LUCAN’s training and scientific‑computing capacity.
Model Purpose and Capabilities
LUCAN is designed to address the long‑standing trial‑and‑error approach in industries such as chemicals, pharmaceuticals, biotechnology, energy and advanced manufacturing. By taking a desired industrial outcome, raw material inputs and factory constraints, the model computationally reasons through molecular behavior, process conditions and manufacturing‑scale physics to propose how the outcome can be achieved, linking discovery, process engineering and manufacturing in a single framework.
Early Industrial Deployments
In a specialty‑chemicals pilot, Shodh AI reports that LUCAN‑driven modelling identified process changes that yielded a 17 % increase in product from the same raw material. The company is also applying the model to biomanufacturing, formulation and broader industrial process optimisation.
Strategic Significance
The launch illustrates a shift from language‑only foundation models to “Physical AI” that can understand and reason about the physical world. With IndiaAI Mission backing, the development positions India to participate in the emerging computing paradigm rather than merely consume it, potentially reshaping the economics of sectors that currently spend large capital on scaling scientific discoveries to reliable industrial processes.
Reference
Technical details are available in the LUCAN technical whitepaper linked in the release.