Applied Computing unveils Orbital 2.0 world model for the energy industry
Applied Computing has unveiled Orbital 2.0, which the company says is built on the first world model developed for energy. The system learns cause-and-effect relationships across the operation of an entire plant, while each customer's model and data remain in their own environment.
- Orbital 2.0 is built on a world model developed for the energy industry.
- The model learns cause-and-effect relationships in plants, including the movement of material and heat.
- Every customer receives a model that runs in their own environment, and the data and models do not leave it.
- Engineers can create dashboards and applications using natural-language instructions.
- Applied Computing says deployment time can be reduced from weeks to days.
Modeling how plants operate
The first version of Orbital learned patterns in plant data and performed predefined tasks. Orbital 2.0, by contrast, is built on a world model that, according to Applied Computing, learns causal relationships.
The model examines how material and heat move through interconnected equipment and how this affects the plant's overall performance. The system is built on a knowledge graph that describes the plant's physical interactions and develops further over time.
Because the model seeks to understand how a plant operates, it is not tied to a single workflow or task. Applied Computing says this allows it to adapt to what a particular engineer needs.
Models that run in each customer's own environment
Every customer receives their own world model, which runs and evolves in their own environment. According to the announcement, customer data and models never leave that environment.
Orbital works with live and historical plant data. It identifies when something is operating incorrectly, attempts to explain the causes based on the physics of the process, and then recommends the next step. The system is designed to operate energy plants at their physical optimum every day.
Applied Computing says Orbital is already delivering results at several of the world's largest energy and petrochemical facilities. Through KBR INSITE®, it is also being used in ammonia and olefin production units.
New features that engineers can customize
One of the new features in Orbital 2.0 is an on-demand dashboard builder. An engineer can describe the interface they want or choose from Orbital's suggestions, and the system creates it from the plant's live data without manual intervention.
Recurring tasks, such as the 8 a.m. shift report or checking heat and material balances, can be turned into applications with a simple instruction. The system also learns from every interaction and from engineers' direct feedback. As a result, the company says, it improves with use, while knowledge about the plant is retained even when employees leave.
Under the new approach, Applied Computing's researchers no longer need to separately tune and adapt the model in every customer's environment. The company says this cuts deployment time from weeks to days and enables deployments to be scaled in parallel. Users can also quickly customize and extend Orbital.
The launch and the company behind it
Applied Computing unveiled Orbital 2.0 on November 3, 2026, at ADIPEC 2026 in Abu Dhabi. According to Callum Adamson, CEO and co-founder, the company's mission is to operate every energy system at its physical optimum, and the fastest way to achieve that is to put Orbital directly in engineers' hands.
“Until now, adapting the model to individual plants depended on our researchers. That barrier is gone: Orbital can now learn autonomously and be fully customized by the engineers who use it every day,” Adamson said.
Applied Computing was founded in London and develops AI systems for energy operations. Orbital combines physics-based intelligence with chemical engineering, time-series forecasting and language models. In July 2026, the company closed a $20 million Series A funding round led by KBR, with participation from Databricks Ventures.
