Orbital AI Promises to Turn Energy Plants Into Real-Time Thinking Machines

Jul 161 views

Key takeaways

Applied Computing raised $20M Series A led by KBR to expand its Orbital foundation AI model for oil, gas, and petrochemical facilities globally.
Orbital combines time-series, physics-based, and language models to flag anomalies and run facility simulations in minutes rather than days.
The startup claims double-digit millions in ARR within 18 months, with a new Houston office and a planned European oil major partnership on the horizon.

Applied Computing, a London-based startup founded in 2023, has closed a $20 million Series A round led by engineering heavyweight KBR, with Databricks Ventures also participating. The company is building what it describes as a specialized foundation AI model — called Orbital — designed specifically for the oil, gas, refining, and petrochemical sectors. The fresh capital will fund international expansion, research and engineering hiring, and deeper deployments with energy clients across North America, Europe, and the Middle East.

The problem Applied Computing is trying to solve is deceptively simple on the surface but staggering in scale. A single energy facility can contain thousands of sensors tracking temperature, pressure, velocity, viscosity, and dozens of other variables at any given moment. Despite all this instrumentation, co-founder and CEO Callum Adamson says operators are actually making decisions based on less than 8% of the data available to them. The bottleneck isn't collection — it's integration. Sensor readings, engineering documentation, and the underlying physics and chemistry of industrial processes rarely talk to each other in real time, and that gap costs operators both efficiency and safety margin.

Orbital is designed to close that gap. Unlike conventional large language models that predict text, Orbital combines a time-series model, a physics-based model, and a language model into a single system capable of predicting the operational state of an entire facility. It can flag anomalies, trace their root causes, and simulate whether a proposed fix might create downstream problems elsewhere in the plant — all within minutes. Adamson claims the product compresses investigations that previously consumed days or weeks into a matter of seconds, reducing energy waste and protecting output levels.

The startup has moved quickly since emerging from stealth. Applied Computing says it has reached double-digit millions in annual recurring revenue within 18 months of launch, with Orbital already deployed at undisclosed large publicly listed upstream and downstream energy companies. Its partners include Indian tech firm Wipro and KBR, which has integrated Orbital into its INSITE 3.0 digital platform and is using it for ammonia production operations. A partnership with a European oil major is also expected to be announced in the coming weeks.

Applied Computing now operates from three locations — its London headquarters, an operational hub in Bengaluru, and a newly opened Houston office that positions the company closer to existing North American customers. The company faces established rivals in the space, including AspenTech, AVEVA, Cognite, and Seeq, but Adamson argues that the true competitive moat lies in AI research talent rather than data access or domain knowledge alone. He contends that top-tier AI researchers are unlikely to flock to legacy energy corporations, giving dedicated AI-native startups a structural hiring advantage.

The bigger picture

Applied Computing's $20 million raise is a meaningful signal that industrial AI — long talked about in vague, aspirational terms — is moving into a phase of serious commercial traction. Reaching double-digit ARR within 18 months, without a splashy public launch, suggests the company found genuine product-market fit before it found its press cycle. That's not common in enterprise AI, where sales cycles are notoriously slow and proof-of-concept projects often die before reaching production. The KBR partnership is particularly notable: having an engineering giant embed your model into its flagship digital platform is the kind of distribution shortcut that most startups would take years to negotiate.

The competitive landscape is genuinely crowded, though. AspenTech and AVEVA have decades of industrial data and deep customer relationships baked into their positioning, while newer entrants like Cognite and Seeq have built real footholds in the data infrastructure layer. Applied Computing's bet is that none of these players have assembled an AI research team capable of building a true multi-modal foundation model for industrial operations — and that this gap is harder to close than it looks. Whether that's correct depends entirely on whether Orbital's architecture can sustain its accuracy edge as rivals respond with their own model investments or acquisitions.

The Middle East expansion angle is worth watching closely. Gulf state energy operators have been aggressively investing in digital transformation and process optimization as they try to extract more value from existing infrastructure. If Applied Computing can land one or two flagship deployments there, it would dramatically accelerate its data flywheel — the proprietary operational data from live facilities being the asset that makes Orbital progressively harder to replicate. The real long-term question is whether the company can scale its AI research team fast enough to stay ahead of well-capitalized incumbents who will almost certainly view this space as strategic.

LagPing's take

We're covering Applied Computing's Series A because it represents something we don't see often enough in energy tech: a company that quietly built real revenue before announcing itself to the world. At LagPing, we track how AI is reshaping industries beyond the usual consumer-facing headlines, and the oil and gas sector — despite its complexity — is one of the most data-rich environments on earth. The fact that operators are only using 8% of available sensor data isn't an abstract inefficiency; it has direct consequences for energy costs, emissions, and safety. Orbital's architecture — fusing time-series, physics, and language models — is genuinely novel compared to what most industrial AI vendors are pitching. We think this story deserves attention from readers who care about where practical, high-stakes AI deployment is actually happening right now, not just where the hype is pointing.

X / Twitter WhatsApp

View at TechCrunch AI

Shop AI & tech on Amazon

As an Amazon Associate, LagPing earns from qualifying purchases. Product links are affiliate links.