Applied Computing raises $20M for Orbital, a plant-scale AI model for oil & gas

Applied Computing, a London startup founded in 2023, has just closed a $20 million Series A led by engineering firm KBR, with participation from Databricks Ventures. The company claims growth from stealth to 'double-digit millions' in annual recurring revenue in under eighteen months, is opening a Houston office and preparing a Middle East expansion — three markers of real traction with major oil and gas operators.

The technological core is Orbital, a foundation model that goes beyond text: it fuses time-series analysis, physics-based modeling and language processing to predict the state of a facility as a whole. Where an isolated sensor triggers a local alert, Orbital reasons under physical constraints and equipment limits, and simulates how a change at one point of the process propagates to every other. CEO Callum Adamson quantifies the untapped value: sites reportedly use 'less than 8%' of available data, and Orbital would compress into seconds investigations that today tie up engineers for days or weeks.

The business case is twofold: cutting energy consumption at constant output, and de-risking operating decisions currently made blind on aging assets. A round led by KBR — rather than a generalist fund — says something about the thesis: process-industry AI is validated first by those who design and run the plants, before it convinces financial investors. Named customers stay anonymous ('large, publicly listed' upstream and downstream players, 'a major U.S. upstream operator'), a sign of a sector still cautious about its AI communications.

Cardan-AI analysis: Orbital illustrates the move from fragmented O&G AI — per-equipment predictive maintenance, per-unit optimization — to a single 'plant-scale' model that brings the operational digital twin closer to real time. For an energy operator, the question is no longer 'do we need sensor AI?' but 'which data, which governance and which industrial-IS integration to get past the 8% we actually use?'. That is exactly the diagnosis Cardan-AI runs with industrial leadership teams: map the dormant data, prioritize fast-ROI use cases, and build the path toward AI-augmented operations without blind dependence on a single vendor.