AI Can Potentially Stem Refinery Closures In Europe, Says Startup Boss

Forbes

15-07-2025

Illustration courtesy Getty Creatives
The situation in Europe's refining and petrochemical sector is nothing short of dire. On average one refinery a month has shutdown over the course of 2025, based on data from across the wider European Economic Area and U.K.
Many others are staring at closures or conversions due to a combination of factors including high energy costs, increased competition from new refineries in Asia and the Middle East, and wider deindustrialization driven by continent-wide carbon neutrality regulations and policy.
Artificial Intelligence can stem the tide and potentially even reverse it, according to a London, U.K.-based energy AI startup boss.
'The prevailing political and cultural narrative of carbon neutrality in Europe is going to see the continent reduce its industrial output by about 35% in the next 15 years, while the demand grows by 60% in the that time frame,' said Callum Adamson, co-founder and CEO of Applied Computing.
'If we don't change things, all that's going to happen is Europe continuing to have higher energy prices, and then – as a second order effect – place the continent at the mercy of nations that are continuing to industrialize. AI premised efficiencies can stop and potentially even reverse this decline before it is too late.'
Fresh from a £9 million ($11.1 billion) seed funding round for his startup in June, Adamson admitted: 'People may label me as politically incorrect and level the charge that I am bigging up energy and industrial AI because it's where our startup's business aspirations lie. For starters, being politically correct rarely protects the future.
"Secondly, the situation in Europe is visibly bleak to the point that very soon we would be left at the mercy of international markets, and are part of the way there already. Thirdly, and unashamedly, I believe the future of the energy business, whether in Europe or beyond, will be both built and secured by AI tools.'
AI Can Change Things
The tool Adamson and his team are actively promoting is Orbital, the startup's flagship product built using multi-foundation AI.
Applied Computing claims it is powered by a 'new class of models built to optimise the physical world'. That's not just language models but also time series, physics and chemical engineering models delivering explainable AI that can be trusted in real-world applications.
Orbital utilizes 100% of available data from downstream energy facilities – compared to 8% captured by traditional methods – and is outperforming previously benchmarked software by 90% in key metrics, according to Adamson.
Applied Computing CEO and Co-founder Callum Anderson (center) at the startup's demo day in London, ... More U.K. on June 10, 2025.
It could potentially reduce energy consumption by 10% across refineries and petrochemical sites resulting in billions saved, he added.
Applied Computing demonstrated Orbital's capabilities last month, using an ABB constructed demo plant in London, U.K., to a select group of industry analysts and sector executives.
'Orbital is one of the few products, and Applied Computing probably the only company in this space that's almost entirely product led. We are completely transparent on benchmarking and product capability.
'What we are seeking to deliver is an understandable, referenceable, retainable lens through which to view AI in an energy industry specific context.'
The startup was formally founded by Adamson and co-founder Sam Tukra, following a year of self-funded independent research, before even a penny of investment flowed in.
'The energy industry doesn't move fast, and we wanted to give it reasons to move fast. Orbital is now referenceable. It works and we remain open to third-party scrutiny.'
The company's raise, led by Stride.VC and Repeat.VC, was among the largest ever for a UK-based AI firm at seed stage. For Applied Computing, this signals growing investor appetite for energy technology that goes beyond renewables to address the traditional energy legacy infrastructure that still powers the world.
'Our ambition is to become category dominant intelligence provider in the energy industry and that means upstream and downstream oil and gas, petrochemicals and liquefied natural gas as well as renewables.
'Of course, our first objective is to scale-up our refinery market penetration because that would be one of our first use cases, and then we move to upstream, LNG and renewables. The plan is to move into a good customer base in refining over the next 12 months. We're currently deployed in two refineries and look to expanding that to ten over the near-term,' Adamson said.
'Refining is where complexity, scale, and impact converge. It's not just hard – it's the apex of validation. If you can build intelligence that works here, it will work anywhere.'
The Journey Ahead
Bearing the challenge in mind, Applied Computing has doubled in size since January with strategic hires from Shell, Palantir, BP Launchpad and Imperial College. It is currently preparing for a Series A funding round later this year.
Adamson said as much as he'd like Orbital and Applied Computing to be part of some profound shift in Europe, bulk of the interest is coming from markets far beyond the continent.
'Europe is walking around in circles because they don't know which way the wind is going to blow next. On the other hand, in Asia, Middle East and pockets of the U.S. we see clear appetite for AI solutions from refinery operators that can deliver significant financial and environmental improvements. For us, whether its a greenfield or brownfield site - we're up for it, as we embark on journey and seek partners along the way.'