Applied Computing opens Bangalore office as India becomes ground zero for next-gen industrialAI

Former Shell AI chief Dan Jeavons on joining Applied Computing, moving to India, and deploying foundation models inside live refineries.

British AI company Applied Computing, which develops foundational AI for energy operators, today announces the opening of its new office in Bangalore, marking its official expansion into India and deepening its commitment to one of the world’s most strategically significant energy markets. The expansion will create new jobs across AI research, engineering, energy modelling and commercial operations.

The move follows significant traction in India, where Orbital has already been proven inside major refining environments and is now being actively deployed with leading operators.

Applied Computing’s flagship platform, Orbital, is the first foundation model built specifically for energy operations, bringing superintelligent, physics-grounded optimisation to some of the most complex industrial environments in the world.

Applied Computing has built-up a senior team in India, including the appointment of former Shell executive Dan Jeavons, one of the world’s leading industrial AI figures, who relocated from London to Bangalore several years ago.

Following his decision to join the firm this summer, Dan decided to remain in India. 

Jeavons previously led Shell’s global AI programme and brings two decades of experience spanning upstream, downstream and integrated gas. I spoke to him to learn more about the country, why it’s been gaining so much traction when earlier startups haven’t, and its market expansion into India.

Jeavons spent almost 20 years with Shell in different capacities, always in the data and analytics space. And then data and analytics blurred into AI as the world headed in that direction.

For the last 13 years, he was leading Shell’s core AI program; his last role at Shell was VP for Computational Science and Digital Innovation. 

He explained, “I led about 350 researchers globally, working on everything from seismic processing to wind turbines, manufacturing plants, and electric vehicle charging—building various types of AI for all of that.”

During this time, he got to know Sam Tukra, “an incredible talent we quickly identified at Shell, but he was very convinced he needed to go and start his own company. So he built what is now Applied Computing, where he now works as Chief AI Officer.”

‘My ChatGPT moment’: why Orbital convinced a Shell veteran to switch sides

Tukra visited him after putting together a research team from Imperial and partnering with Callum Adamson, the CEO and co-founder, and an Entrepreneur-in-Residence at Imperial.

He recounts:

“He came to see me and showed me what Orbital is, the foundation model we’re developing at Applied Computing.

I always say it was my own personal ChatGPT moment. 

I’d been very close to the developments of large language models, so ChatGPT itself didn’t really surprise me.”

The real breakthrough he saw in Orbital was the ability to combine physics, time series and language into a common foundation model. 

“For our world — the world of energy operations — that is an absolute game-changer because it brings into one integrated model almost every question you could ask about the operations of a site.

So it becomes a true general AI you can deploy into some of the world’s most complex industrial landscapes.

I got very excited, and long story short, they convinced me to come over and help build the company. It wasn’t in the life plan — but here I am as President.”

Why industrial AI struggled — and why that’s now changing

I’ve written 100s of thousands of words about industrial IoT throughout my career. However, Industrial IoT, in many respects, failed to deliver on its promise.

Most factories are built on decades-old machinery, proprietary protocols, and safety-critical systems that are difficult or costly to connect, making integration economics unattractive. Projects also suffered from unclear and slow ROI: large upfront investment in sensors, connectivity, integration, and security often delivered only incremental savings, so pilots rarely scaled.

Further, while IIoT generated vast amounts of data, organisations lacked the analytics maturity to turn that data into actionable outcomes as they depended on an AI capability that didn’t yet exist. Only now, with foundation models, edge AI, and domain-specific intelligence, is it becoming feasible to rethink industrial systems.

Jeavons admits that, as someone who spent a decade working on Industry 4.0 projects, data-driven methods only really impacted the peripheral operations.

“You could do equipment failure prediction, integrity management, inspection, rust detection, classification — those worked and delivered value.

But the utopia we are glimpsing now with AI? I felt that was achievable with the prior generation of tech. But it wasn’t.

It wasn’t explainable enough. It was too black box. It wasn’t appropriate to deploy into the core of operations. With the next generation of models, all of that is changing. And I think we will see a radical transformation in heavy industry in the next few years.”

According to Jeavons, most critical infrastructure runs on physics-based simulations. When you design a facility, these are the equations that govern its operation, because they’re governed by the laws of physics. 

“We can say that if we operate within these constraints and boundaries, we can produce the desired output from the process. We run that process a lot of times, then embed that into a control system.

What that allows you to do is run the facility in steady state within those boundary conditions. Operators are there to make sure those operating limits aren’t breached.”

However, this fails to integrate all the data the plant generates continuously. That data gets used, but only for root-cause analysis or incident detection, but its siloed and disconnected from operations. Then you have a whole variety of engineering disciplines sitting around the plant using subsets of that data to derive insights that operators might want to know.

Orbital completely rethinks this. It can combine the physics from the simulator and integrate it with the data — not just time-series data, but also language data: the reports written, the work orders generated, the inspection reports created five years ago. Jeavons equates it with the aeroplane and the control tower. 

“The pilots are flying it — that’s what operators do. But in these sites, you don’t have the control tower where you can see everything else going on. We’re saying: you can build a control tower.

Across not just one site, but 40 sites. You can compare operations. You can look at when a piece of machinery failed and ask: where else is that likely to happen next? You can say: I’ve seen this integrity condition here, under these conditions—have we checked over there? Your ability to look across the entire business and empower that with A I— that’s where the transformation comes from.”

Asking operational questions in natural language while protecting proprietary data 

For Jeavons, the language interface makes an enormous difference:

“The ability to ask simple questions in natural language and interrogate all the data — engineering drawings, shift logs, maintenance history, time series — that ability to cut across silos is game-changing.

Each engineering discipline is looking at one slice. Orbital lets you ask questions across all of them.”

I was curious about how Orbital handles proprietary data, a major issue in industrial environments. Crucially, Applied Computing brings its model to the customer’s data. Orbital has a foundational understanding of physics, the domain language, and time-series behaviour. 

“It’s a transfer-learning principle. We augment it with the customer’s data—in their environment,” explained Jeavons.

“Training happens inside their infrastructure, or an environment they control. That means we don’t expose their data to the world or to other customers.

We do not ask for their data to train a global model. The model remains our IP, but everything produced in the customer’s environment is theirs to use under the contractual terms. And the data never leaves their environment.”

From pilot to production in under a year

Less than a year into the market, the company is deploying into real customer environments and seeing strong outcomes. 

According to Jeavons, “The biggest thing customers tell us is the ability to approach a problem from multiple different angles, unlike before, where you needed a whole team of experts. The biggest benefit is the ability to use AI to rethink how you run your business and answer questions you couldn’t answer before. That’s where senior leaders are really engaging—because they believe this will change the game.”

Applied Computing raised £9 million in May this year. The funds are being used for research: “At our core, we are a research company developing a next-generation foundation model for the energy industry,” explains Jeavons.

“AI moves daily, and we must stay at the cutting edge.”

The second use is go-to-market strategy. The company has hired domain experts as “it’s great having a killer model. The challenge is deploying it, solving users’ problems, and earning the right to expand within accounts,” shared Jeavons. 

He sees his role as to bridge the tech with industry needs and shape the narrative and deployment to drive impact. 

“To drive material change in these organisations, you have to work at the C-suite level. It’s not just technology—it’s rethinking how you run your business.

That’s why we focus on a few strategic customers and high-level conversations to drive outcomes.”  

Rising energy demand meets India’s leapfrog-ready infrastructure

Applied Computing’s expansion comes as India’s energy landscape reaches a pivotal moment.

While global policy trends push towards decarbonisation, India’s energy demand continues to rise sharply, driven by industrial growth and a rapidly expanding population.  According to Jeavons, India is the fastest-growing energy economy in the world.

“There’s incredible talent here. I lived in Bangalore for three years — I moved here with Shell and stayed because I’ve been so taken with the country and its potential. We’re celebrating the opening with friends, customers, and partners. We’ve already had customers come into the office to work with us and see the research team. It creates a phenomenal space for accelerating the technology.”

He believes that. while Europe has been phenomenal in many areas— “just look at the startups that have emerged from the region.” India can leapfrog in areas where it isn’t constrained by legacy systems.

“That entrepreneurial instinct is very real. That’s why I’m excited about Applied Computing being here: we can run the leapfrog playbook.”

He shared: 

“Orbital is already delivering results with customers here, including at some of the largest refineries on the planet.

By establishing our base in Bangalore, we’re investing directly in the talent, partners and ecosystem that will define the future of industrial intelligence.”

It’s a sentiment echoed throughout the company. According to Callum Adamson, CEO and co-founder of Applied Computing,  “India is not just another geography for Applied Computing – it is our primary market and a proving ground for the future of industrial AI.”

“The country’s refining and petrochemical sectors are central to the global economy, and the decisions made here will influence energy stability and emissions worldwide. 

Orbital is already deployed in India, delivering impact at unprecedented scale, and our investment in Bangalore strengthens our ability to support operators as they modernise and transform their most critical infrastructure.”

Why India is becoming ground zero for industrial AI

India’s country’s refining and petrochemical sectors are set to grow substantially over the next decade, and many of its critical assets rely on ageing infrastructure where AI-driven optimisation can have outsized impact.

India’s openness to technology adoption and its willingness to deploy AI at an operational scale make it one of the most important markets globally for industrial intelligence. 

The combination of demand growth, infrastructure complexity and a culture of technological experimentation positions India as the ideal environment for Orbital to deliver immediate system-wide benefits.

Jeavons is joined by Hari Ramani, Vice President of Commercial Markets, who will lead customer engagement and global market development:

“Energy operators in India are managing extraordinary complexity across ageing and emerging infrastructure. They’re looking for solutions that improve efficiency today while preparing for a more sustainable tomorrow. 

Orbital provides that bridge — delivering actionable, physics-grounded intelligence across entire facilities. The appetite for real-world AI adoption here is unmatched, and today’s expansion positions us to serve this demand at scale.”

Lead image: Dan Jeavons, Applied Computing.