The car that existed before it existed
In four vehicle development programs, across engines manufactured by Jaguar Land Rover, Isuzu, Mahindra and Sharda Motor Industries, engineers using the same simulation software cut development time by an average of 58% and reduced costs by 43%. They are not projections. Those are the results – documented in production programs, on hardware that’s on the road today.
These numbers are worth a moment because they raise the question of what the automotive industry generally regards as engineering. It’s a question whether a manufacturer can’t do this, especially in India, where development timelines are shrinking and powertrain complexity is increasing manifold.
When Mahindra launched the BE 6 and XEV 9e last year, the vehicles had already survived the heat of Rajasthan, climbed the Sahyadris and traveled on the Mumbai-Pune Expressway hundreds of times. Long before the first physical prototypes were signed off, the company’s engineering teams had virtually run each of those scenarios – testing how Indian heat, Indian gradients and Indian traffic would affect range, battery performance and thermal behaviour.
They then used those findings to shape the design. By the time the mule hit the tarmac, it was not aware of the problems. It was confirming that a design already stress-tested in a way that no proving ground could replicate at speed or scale – held up. Mahindra’s previous program to develop BS6 engines through the same platform came in at 50% of its original development timeline and 40% below the estimated cost.
This is system simulation in practice.
What is simulation, and how is it different from normal CAD?
The automotive industry is familiar with CAD – computer-aided design – which creates the geometry of a part or vehicle. Simulation is a separate discipline. Computer-aided engineering (CAE) tools model not just how something looks, but how it behaves under thermal stress, under electrical load, under the particular punishment of Indian roads, and in thousands of operating conditions simultaneously.
In the vehicle context, this means creating mathematical models of individual systems, battery packs, thermal circuits, engines, HVAC, and integrating them so that software can solve their interactions together. What effect does that have on the cabin temperature when the battery is working hard on a 42-degree afternoon? How does this change the limit calculation? How should the cooling circuit respond, and what is the energy cost of that response?
These are not questions that a CAD file can answer. A simulation model – in minutes, not months, and without consuming a single physical part.
As Matthew Warner, vice president of Gamma Technologies, said at his company’s annual technical conference in Pune this February: “The idea is to reuse the value that exists in your CAE model to answer questions in other parts of the development organization – control, testing, requirements engineering – to create a model that delivers consistent results regardless of the application.”
One source of truth, used everywhere. The 58% time savings resides, in large part, in that sentence.
Why such a hurry?
Three forces have converged to make simulation less of an engineering priority and more of a business necessity.
The first is the powertrain complication. The neat narrative of the green mobility transition has yielded some pitfalls. OEMs are simultaneously developing battery EVs, hybrids and updated internal combustion vehicles – not sequentially but in parallel with engineering teams. Each architecture demands its own simulation environment and its own verification loops. The workload has increased manifold without proportionate expansion in the number of employees or time-frame.
The second is the shrinking product cycle. “The product development cycle has reduced significantly, from 3 to 5 years to about 2 years now,” said Dr NH Walke, senior director, ARAI, at the SIAT 2026 conclave in Pune. “Now it’s concurrent engineering. Simulation, component development, and proving – all this testing has to happen simultaneously.”
A program that could discover problems in a timely manner must now step in and solve them in parallel. Isuzu Technical’s commercial diesel program, using the same simulation platform, reduced dynamometer testing hours by 60% and costs by 45% in time that would otherwise have been spent in physical test queues. Jaguar Land Rover’s Ingenium 2.0L diesel program achieved a 50% reduction in development cycle time and 35–40% cost savings.
This pattern is consistent enough across programs to be structural rather than exceptional.
The third strength is specifically Indian: the historic cost of proving vehicles overseas. For verification infrastructure, which barely existed domestically, Indian OEMs traveled to European facilities – absorbing the time, cost and strategic inconvenience of developing products calibrated in foreign conditions.
ARAI is closing the gap with domestic investments in crash labs, advanced battery testing facilities and ADAS proving grounds. The simulation actually increases the value of that infrastructure by reducing the number of physical runs required for a valid design – which, in practice, also reduces how many flights the Indian engineering team needs to book to Stuttgart or Gedeon.
The market has priced the trajectory accordingly. According to Precedence Research, automotive simulation software was valued at $7.06 billion globally in 2025 and is projected to reach $24.35 billion by 2034, a CAGR of 14.75%.
A company that saw it coming in 1994
Gamma Technologies was founded thirty-one years ago in Illinois on a premise that was counterintuitive at the time. Most CAE vendors were creating tools for single-physics, single-component analysis. GT’s founders believed that the more valuable problem was the system as a whole.
“The products don’t exist in a single physics domain – they exist as multi-physics systems,” says Dimple Shah, who has led the company as CEO since 2020 and worked at CAE since 1991.
Thirty years later, that thesis is mainstream. GT-Suite, Gamma’s flagship platform, covers engine performance, battery electrochemistry, electric powertrain, thermal management, exhaust treatment, fuel cell and HVAC in an integrated environment. Its user base spans every major global OEM; In India, this includes Tata Motors and Mahindra.
GTTC 2026 was the opportunity to launch GT Intelligence Studio – an AI-native addition to the platform that layers generative AI and machine-learning meta-models onto the existing physics engine. The announcement raises an obvious question among engineers working in safety-critical systems: How do you trust an AI model when the output affects a brake calibration or battery management decision?
Shah’s response is grassroots rather than promotional. “The quality of meta-models depends on the quality of the datasets they are trained on. Our meta-models are trained on physics data, which can also be augmented by data from external sources. Because a large part comes from physics itself, confidence in the model is high. One should not blindly deploy a meta-model; like any simulation model, you need confidence that it is applicable in the range you want to deploy it.”
Warner adds the context that matters: Machine learning components are not a new development. “This has been in our software for about 20 years. ML meta-models are often not deployed on an actual vehicle in operation. They are used within the development process of the vehicle, so in that use case they are not safety mission-critical.”
What is really new is the generative AI layer and, more practically, the cloud platform GT-Play that puts validated simulation models in the hands of engineers in an enterprise who are not simulation experts, but rather control teams, test planners, and product managers.
With an honest warning, India is on par
The Mahindra presentation at GTTC 2026 showed what the Indian engineering team working on the ground looks like in practice.
On the BE6 and Aerodynamic trade-off analysis mapped the relationships between drag coefficient, vehicle weight and range in various feature configurations before any of the physical prototypes were carried out.
In a pan-India virtual drive exercise, highways, ghats and cities were mapped across geographical regions of the country, taking into account gradients, AC load and regional speed distribution. The same team designed the simulations to incorporate virtual calibration on the front end, battery health monitoring and predictive maintenance after launch, and digital twins that give engineers objective data against which to audit subjective customer feedback from the field.
Shah’s assessment of where Indian OEMs stand is spot on: “Indian companies are really aspirational. They are becoming very strong contenders globally. At this conference, when I compare the quality of papers being presented by our community with our European, American and Japanese conferences, I don’t see any difference. The user community is at par.”
Whatever warning he gives, it is better to leave it. “Many global OEM companies focus on five to ten year technologies while building core expertise in-house. Some companies in India can do this and more. One area where India can do more is fundamental research. Today, they do this, but largely through partners and collaborators rather than in-house. This is going to be one of the key trends in the coming years.”
There is a clear difference between using simulation tools at a world-class level and generating the fundamental knowledge that dictates what these tools will do next. The 58% and 43% figures were earned by engineering teams that knew how to run the software. The companies that will define the next version of that software will be the ones building that knowledge ahead of the programs that will need it.
