By 
By: Karnvir Mundrey
An IFCCI site visit to Dassault Systèmes in Bengaluru revealed how cars, aircraft, factories and even cities are now built twice-and raised a question India’s manufacturers can no longer avoid.
“Dassault makes the Rafale,” I said to my wife before leaving home. “So what exactly are they doing in Jayanagar? Where’s the runway?”
It was meant as a joke. It turned out to be the right question, asked for the wrong reason.
Jayanagar is a neighbourhood of idli, dosa and old South Indian restaurants, as our hosts cheerfully admitted. There is no hangar and no airstrip. What there is, on land that was once industrial, is a room where the same software that shapes a Boeing airliner also shapes a Procter & Gamble shampoo bottle.
On 11 September 2026, the Indo-French Chamber of Commerce and Industry brought its members to Dassault Systèmes for a site visit on virtual simulation and industrial innovation.
Here is what I learned. Dassault Systèmes has a hand in a surprising share of the physical things around you, yet most people have never heard of it. That invisibility hides a story that matters enormously for India: a story about time, trust, and who owns industrial knowledge once artificial intelligence starts learning from it.
So this is not really a story about a software company. It is about a new kind of runway, the place where the future takes off before it exists, and whether India intends to use it.
I went looking for a runway. By the end of the morning, I realised I had been standing on one.
Deepak NG, Managing Director India, Dassault Systems
A landmark they decided to build
Mr. Deepak NG, Managing Director of Dassault Systèmes India, began by explaining the address. The company used to operate from the Leela Palace, which had room for only about 80 people. It now employs close to 1,200 in Bengaluru. When colleagues grumbled about leaving a famous address, he told them to go and create a landmark rather than depend on somebody else’s.
He had a precedent. Bengaluru already has a DELMIA Circle, named after the company’s manufacturing-software brand, whose R&D centre stood nearby from 1996 to 2013. Dassault is betting Jayanagar will be next.
The most famous company you have never heard of
The first surprise was a trivia question. Dassault Systèmes was spun out of Dassault Aviation in 1981 to build the design tool that became CATIA. So who was its first customer?
Not an aircraft maker. Honda.
The second surprise was how long the company stayed in the shadows. For decades, IBM sold its software worldwide while Dassault built the technology in the background. Only around 2008 did Dassault take its business back. Today it serves 390,000 customers, employs roughly 25,000 people, and reported 2025 revenue of €6.24 billion. Few people outside engineering could tell you what it does.
“Our brands are more powerful than us,” one executive told the room. Engineers know CATIA, designers know SOLIDWORKS, factory planners know DELMIA, and scientists know BIOVIA. The parent company mostly draws blank looks, or questions about fighter jets. The same executive admitted that his own title includes “global affairs”, though nobody remembers why. There is nothing global about it, he joked, and no affair either.
I have run a communications firm since 2006, and almost every B2B client I have advised treats invisibility as a problem to fix. Dassault suggests a more complicated truth. For a platform company, invisibility can be a moat. Engineers do not think they are buying from Dassault; they think they are working in CATIA. That is a deeper loyalty than brand awareness can buy.
It is a clever position.
Crash it before you build it
The simplest way to understand Dassault Systèmes is through one question:
What if we could make our mistakes in a virtual world before paying for them in the real one?
For most of industrial history, making things followed an exhausting loop. Draw. Build. Test. Fail. Fix. Repeat. Every lap cost material, money and months.
Simulation broke the loop. A car door can now be slammed a million times in software before a hinge exists. A production line can run virtually before a machine is installed. Aerospace engineers call the ideal “right first time”: the first physical prototype is expected to fly.
Dassault calls this a virtual twin: not a static 3D picture, but a living digital copy connected to the behaviour, processes and data of the real thing. The real prize is a continuous digital thread from first sketch to final retirement.
It sounds like an engineering story. It is really a story about time.
The new currency is months
A new passenger-vehicle programme in India, the Dassault team said, may take roughly 36 to 42 months from concept to production. Some Chinese manufacturers work on cycles closer to 18. In its work with Mahindra, the stated ambition is to cut about 42 months to about 30 by connecting design, validation, sourcing and manufacturing more tightly.
We have long measured Indian competitiveness in wages, talent and market size. The new contest is measured in months. The country that learns fastest may beat the country that manufactures cheapest.
My training in finance at the London School of Economics made me notice another angle. Analysts usually learn about a new platform when capital expenditure is announced. By then, the programme has often lived in software for years, with its design frozen, suppliers shortlisted and production line rehearsed. Industrial software is the earliest map of where corporate investment is heading.
And some of the most important of those maps have been drawn, for years, in Bengaluru.
Bengaluru’s open secret
Every major Indian IT services company, our host explained, has a second business outsiders rarely notice: engineering services. For two decades, Indian engineers inside these firms have designed trains, aircraft structures and engine components for Western manufacturers, largely in French software.
These began as offshore development centres. Manufacturers later took them in-house. Today we call them Global Capability Centres, as though they were new. They are not. When a new widebody aircraft is coming, our host remarked, people in Bengaluru know years ahead, because they are already working on it.
India’s aerospace ambitions are usually framed around factories. But the design layer came first, and it has quietly been one of India’s strongest links in global aerospace.
Which is why I could not resist my next question.
So I asked about Bengaluru
“If you can build a virtual twin of a car, an aircraft carrier or a factory,” I asked our hosts, “why not Bengaluru? Why does a city full of technologists remain so chaotic?”
The answer was uncomfortable: it can.
Dassault has worked on virtual-city projects including Singapore, and in response the executives described Indian work involving Jaipur, Amaravati and parts of Bengaluru. A city twin can simulate flooding after an extreme downpour, traffic from a proposed flyover, heat between buildings, evacuation routes, and the utilities hidden underground. It can even show how mature trees could be moved instead of felled.
Then came the detail that made the room lean in. Combine municipal records with 3D building data, they explained, and the model could reveal differences between the floors declared for property tax and the floors that actually exist. A building paying tax on two floors might visibly have three.
The twin would not just display the city. It would expose the gap between the city on paper and the city in reality.
That is where technology becomes political. A simulation can show which neighbourhood will flood before a drain is dug. It cannot force a public institution to act on what it reveals. The technology, one speaker replied, may be the easy part. Putting it into practice is the real challenge.
The question is no longer whether such tools exist.
It is whether our institutions genuinely want to see.
An AI that understands gravity
Uma Shankar leads data science and AI for Dassault’s Indian operation. He confessed that if he goes on holiday, something in AI has changed by the time he returns.
His argument was refreshingly simple. A general-purpose language model is fluent but knows no physics. Ask it to apply a force to a fan and it has no idea what inertia is. In a chatbot, that is an embarrassment. In an aircraft, it is a disaster.
His team builds industrial AI in three layers:
- Knowledge. Decades of industry expertise are combined with each customer’s own product data, organised as a knowledge graph.
- Physics. Simulation keeps answers consistent with physical law.
- Reasoning. The system can tell an engineer something no chatbot will: that a design cannot be manufactured on their shop floor.
The result is three Virtual Companions: AURA for business and programme management, LEO for engineering, and MARIE for science. Shankar showed the difference with one question: how do I design a composite bracket? AURA lists past projects that used composites. LEO talks about ply counts and orientations. MARIE goes straight into resin chemistry.
Same question, three kinds of expert.
In the demonstration, a manager asks for the commissioning status of a virtual plant. The system colour-codes the 3D model, notices a supplier has not delivered a heat exchanger, finds a similar unit in another division, and suggests borrowing it.
The most telling proof point was a customer that barely uses Dassault’s core tools. L&T Hydrocarbon uses neither CATIA nor DELMIA; it uses the platform to orchestrate 28 separate systems across its projects.
Impressive.
The pollinating bee
Earlier, explaining how a company that makes nothing physical came to understand so many industries, Mr. Deepak NG had offered an image.
“We are like the pollinating bee,” he said. Dassault does not know how to make cars. It learns from those who do, builds that learning into its tools, and makes it available to everyone. Win the top ten companies in each industry, and their suppliers follow.
Indian readers may know an older version of that image. In the Bhagavata Purana, a wandering sage who learns from twenty-four teachers counts the honeybee among them. The bee takes a little from many flowers and harms none. The same text adds a warning: the bee that hoards its honey loses it to the honey-gatherer.
Both lessons were in the room. Dassault’s customers compete on proprietary know-how, not common knowledge. If the AI learns only from public knowledge, it is generic. If it learns from private knowledge, what stops the bee from carrying Boeing’s pollen to Airbus?
The answer was careful. Each customer’s data and specialised models stay inside that customer’s own cloud environment and are never used to train anything shared. If a licence lapses, the model waits there, dormant, until renewal. Dasault added a candid admission: across roughly 25,000 Indian customers, Dassault has no idea what most of them do with its software unless it is hired as a consultant.
That is the right architecture. But pollination has always been the business model. The line between “improving the common platform” and “carrying one customer’s advantage to another” is drawn in contracts and enforced through audits.
In industry, an answer must be correct, traceable and certifiable. So must the promises about whose knowledge produced it.
From car doors to jawbones
Since 2016, the Bengaluru team has worked with NIMHANS psychiatrist Dr G. Venkatasubramanian, turning brain scans into physics-based models to predict how non-invasive stimulation for schizophrenia will act before it reaches a patient. A similar idea for Parkinson’s tremors was deliberately stopped at simulation because of the ethical questions it raised.
Another project tested personalised jaw implants for cancer patients virtually against years of chewing, heat and electromagnetic exposure.
Car door or jawbone, the principle holds: simulate first; intervene later.
India can make it. Can it prove it?
For me, this was the heart of the visit.
Only days earlier, after a supply-chain event in Vietnam, I had written that India can make anything—but cannot always prove it. In aerospace, that gap becomes brutal. Global manufacturers are drawn to India’s costs, talent and supplier base, yet they keep hitting the same wall: trust in quality, documentation and process discipline.
In aerospace, a millimetre is not a minor disagreement. And no supplier is credible just because its founder swears the factory is capable.
Capability has to become evidence.
The digital thread could become India’s missing trust infrastructure. Connect a part’s design, material, simulation, inspection and manufacturing history, and quality becomes easy to demonstrate while deviations become hard to hide. And when the top manufacturers standardise on one platform, suppliers at every tier must eventually work within it. For an Indian MSME, digital capability is not something to buy after the first aerospace order. It is the entry ticket.
Is the ticket affordable? Asked about pricing, one executive did not flinch: like French wine, he said, it is expensive.
Yet the tools are not reserved for giants. A 40-person eVTOL startup incubated at IIT Madras uses the same technology as Boeing; only the number of seats differs. Dassault’s published startup offer is free to start for companies below €1 million in funding or revenue. Its 3DEXPERIENCE Lab takes no equity, and founders keep their IP. In Tamil Nadu, TANCAM, a centre of excellence set up with TIDCO in 2021, gives MSMEs and startups access to design and simulation tools.
Karnataka, home to so much of India’s aerospace design talent, would benefit from something similar.
So, where is the runway?
Which brings me back to my wife’s question.
Dassault Systèmes does have a runway. It just isn’t made of tarmac.
It is the virtual twin: the place where a car is crashed before it is built, a city floods before the rain arrives, and an implant is chewed on for years before it enters a patient.
The world is starting to exist twice. First as a space where we can fail safely and learn cheaply. Then as the physical reality where we must live with our decisions.
India’s problem has never been a shortage of engineers or ambition. It is our stubborn habit of building first, discovering the conflict later, and repairing forever. The runway offers the opposite discipline:
Anticipate. Test. Collaborate. Prove.
I went to Jayanagar looking for a runway. I came home knowing where it was.
The only question left is how much of India’s future we are willing to test there, before we build it.
Karnvir Mundrey is a narrative strategist and media entrepreneur who helps founders, institutions and international businesses turn complex ideas into influential public stories.
He is the Founder of Atharva Lifesciences Consulting Pvt. Ltd. , Atharva Marcom and Founder Editor of TheFutureOfPR.com. He has also authored a book on Nutraceuticals (available on Amazon). He is also recognized as India’s longest running podcast host, continuously running since 2006!
Reach out at tfofpr@gmail.com or at +918296303806.
Subscribe to TheFutureOfPR.com to get great ideas on life, education, health & fitness, real estate, glamour, jewelry, movies, and podcasts! Follow TheFutureOfPR.com on Facebook , Twitter and Linkedin!
Karnvir Mundrey is also the producer and host of 4 YouTube channels. Finest Fintalk brings you the latest in Finance, LitInMin for Books, The Health Tips Podcast for health and Atharva Marcom for leadership talks
Share this article with people who you think might benefit. They will thank you for it!
Comments