Orbital 2.0

Orbital 2.0 (Applied Computing's next-generation Orbital) launched at ADIPEC on November 3, 2026. Where the first Orbital learned data patterns for fixed tasks, the new stack is rebuilt on an energy world model: a knowledge graph of physical plant interactions that learns how mass and heat move through equipment so the system can explain cause and effect—not just memorize correlations.

Each customer's model runs and evolves inside their own environment; data and weights do not leave. The launch follows a $20M Series A (July 2026) led by KBR with Databricks Ventures. Orbital is already used on large refineries and, via KBR INSITE, on ammonia and olefins units.

What is new

• Dashboards on demand — engineers describe (or accept suggestions for) live boards Orbital builds from plant data.
• Self-running apps — recurring work such as shift reports or heat-and-mass balances becomes automated apps.
• Self-learning — feedback from everyday use updates the model so vendor researchers are less of a bottleneck.

Positioning

Industrial real-time optimization and anomaly explanation for refining, petrochemicals, LNG and related energy systems—physics-grounded and engineer-customizable, not a general office agent.

Who it is for

• Energy operators and EPC/partners (e.g. KBR) that want frontline engineers to extend the AI themselves.
• Teams that require private deployment with data residency on site.

Getting started

1. Read the newsroom post and PR wire.
2. Book a technical briefing and map DCS / historian interfaces.
3. Pilot on one unit (e.g. balance check or shift report) before closing an optimization loop.