SciFM 2026 focuses on the next phase of scientific AI: moving from models that analyze the physical world to systems that actively engage with it.

Recent advances in scientific foundation models and agentic AI have transformed representation learning, multimodal integration, and workflow automation. The emerging challenge is to embed these capabilities directly into laboratories, instruments, facilities, and field systems—creating closed-loop scientific workflows that operate in real time.

SciFM 2026 centers on the integration of:
 

• Foundation models trained on simulation, experimental, and observational data

• Agentic systems capable of planning and multi-step execution

• Robotics and instrument control

• HPC and facility-scale infrastructure

• Safety, validation, and standards for physical AI

The core question is not simply how to predict physical processes, but how to design AI systems that can safely and reliably interact with them.

Key themes include instrumentation intelligence, multimodal scientific foundation models, physics-informed learning, embodied scientific agents, facility-scale autonomy, and benchmarking for real-world deployment.

SciFM 2026 convenes researchers, facility leaders, industry partners, and funding agencies to define the systems architecture required for scalable, interoperable, and trustworthy physically grounded AI.

As scientific AI shifts from digital inference to physical engagement, the technical, institutional, and safety implications expand accordingly. SciFM 2026 provides a focused venue to shape that transition.