Session

Building a Foundation Model for Energy on Databricks with Applied Computing

Overview

ExperienceIn Person
Energy is arguably the most critical industry on earth, and yet the complexity of the processes and the implications of getting it wrong have been barriers to AI adoption, until now. This session will give context to the current challenges faced by the energy sector and explain how Orbital, the foundation model for energy developed by Applied Computing, can address those challenges by providing an AI model that is both physics-informed and domain specific. The session will also unpack how Databricks provides the enterprise-ready backbone for Orbital, supporting rapid data ingestion, creation, indexing, model training and more

Session Speakers

Speaker placeholderIMAGE COMING SOON

Bryce Bartmann

/Head of Forward Deployed Engineering
Applied Computing

Full Summary

Teaching machines to speak physics: inside Applied Computing's Orbital foundation model

Digital AI has spent decades optimizing clicks and conversations, yet the industrial systems that generate energy, materials, and power remain largely under-optimized. Applied Computing argues the biggest untapped dataset on Earth is operational sensor data from plants, grids, and refineries, and unlocking it requires a physics-informed foundation model built on the modern Databricks Platform. The stakes are enormous: a 1 percent efficiency gain in the physical economy implies roughly 350 billion dollars annually, 380 million tons of avoided emissions, and progress on energy access for hundreds of millions of people.

FAQ


It is a model whose outputs are continuously checked against relevant physical laws. Orbital combines a domain-tuned LLM, time-series analytics, a physics model, retrieval over unstructured documents, and a validation layer that blocks recommendations inconsistent with the underlying physics.

Every recommendation passes through a validation layer that cross-checks against the physics model. MLflow evaluations and experiments are used to ensure updates continue to return deterministic, grounded answers acceptable to industrial engineers.

Each customer runs in a dedicated Databricks workspace. Any models trained for a customer are stored in that customer's Unity Catalog, and Applied Computing does not export or fold customer data into a shared world model.

Key industrial protocols and historians, such as OPC DA, OSIsoft PI, Yokogawa ExaQuantum, and Honeywell PHD, lack cloud-native connectors and often run on legacy Windows systems. Custom connectors built with Spark declarative pipelines and Python data sources ingest these streams reliably on premises or in the cloud.

Yes in this case. Because training workloads are bursty, serverless GPUs in Databricks AI Runtime, including A10s and H100s, reduced effective cost and removed the need to negotiate dedicated capacity with hyperscalers.