Open the bonnet of a car from the early 1970s and the story is almost charming. A mechanical engine, a carburettor, a distributor, and not a single computer in sight. Fast forward to today and a typical vehicle carries more than 100 million lines of code. Somewhere between these two generations, the industry quietly changed what it sells.

McKinsey expects the global automotive software and electronics market to reach about $462 billion by 2030. NITI Aayog puts software at roughly 2% of a vehicle's value today, rising to 4 to 5% by 2030, within an automotive software market of around $80 billion.

How the car got a brain

For decades, engines were purely mechanical. Then emission rules and the push for fuel efficiency made precise fuel and ignition control essential, and the engine control unit arrived. One computer became several, and several needed a common language. Bosch's CAN bus, developed in the mid 1980s, gave them one. The US made on-board diagnostics (OBD-II) mandatory in 1996, Europe followed around 2001, and by the early to mid 2000s telematics was a standard bolt-on layer. Plug in a reader, pull fault codes, track a fleet.

Every new feature meant another box, another supplier, another wiring harness. The software-defined vehicle (SDV) flips the logic. A few powerful domain and central computers run the car, and features arrive through over-the-air updates long after it leaves the showroom. Tesla taught the world to expect that. Everyone else is now racing to catch up.

Who is building it

Globally, the stack is being carved up in interesting ways. Qualcomm and NVIDIA supply the compute. Bosch and Continental, once hardware suppliers, now sell software. Volkswagen Group signed a letter of intent with Qualcomm for next-generation driving experiences, while Mercedes-Benz builds its own MB.OS.

Two names deserve more than a passing mention, because they sit at opposite ends of the strategy spectrum and both sell cars in India. Hyundai Motor Group launched Pleos as a software brand in 2025, spanning high-performance vehicle chips and controllers, a connected vehicle operating system, cloud infrastructure and fleet logistics software. The OS uses a zonal architecture with a central vehicle computer. Its infotainment layer, Pleos Connect, is meant to power more than 20 million Hyundai, Kia and Genesis cars by 2030.

BYD is the other pole, and it is vertically integrated in a way few Western carmakers are. Its God's Eye driver-assistance suite and DiLink cockpit run its cars in China. In May 2026 it unveiled Xuanji A3, an in-house 4nm smart-driving chip, reportedly due on a Denza model in 2027, and it says its chip design work goes back to 2002. Until now, God's Eye leaned on NVIDIA and Horizon Robotics silicon, with Momenta and Huawei supporting the algorithms.

The India version is a very different story. BYD sells only battery-electric cars here, the Atto 3, Seal, eMAX 7 and Sealion 7, priced at roughly ₹25 lakh to ₹45 lakh as of late 2025, and one industry report says it raised prices three times in 2026 on forex pressure.

The lesson for India is uncomfortable but useful. In both cases the software architecture is decided in Seoul or Shenzhen, and it arrives with the platform or the car. Tata and Mahindra are not only competing on price. They are competing against stacks backed by decades of platform and silicon investment.

India's contribution is easy to underestimate. NITI Aayog notes that TCS, Infosys and Wipro already supply software for connected vehicles and infotainment, and engineering houses such as KPIT, Tata Elxsi and Tata Technologies work inside many global OEM programmes. A good share of the code in a European car has quite possibly passed through a desk in Pune or Bengaluru.

What EVs really change

People often treat EVs and SDVs as the same story. They are cousins, not twins, but EVs make the software argument far stronger. NITI Aayog estimates that around 190 combustion components disappear in an electric vehicle, and EV-specific parts such as motors, batteries and electronics make up about half of its value. Range, charging speed, thermal behaviour and battery health are all tuned in software, so an update can genuinely make an EV better. Semiconductor content per car is expected to double from $600 in 2022 to $1,200 by 2030. And an EV is a clean sheet. With no legacy engine architecture to protect, central computing is much easier to justify.

Figure 1. Software and chips are claiming a bigger share of every vehicle. Sources: NITI Aayog

The infotainment puzzle

Here is a complaint I keep hearing first-hand from customers. Mercedes, BMW and Audi invest heavily in proprietary infotainment stacks, and many owners feel the everyday experience pays for it. To be fair, these brands do support Apple CarPlay and Android Auto in most models. But the native system remains the centrepiece, and Mercedes now describes MB.OS as its own operating system for the whole car. The logic is easy to see. Control the screen and you control the data, the services and the revenue.

The trouble is a mismatch of clocks. A phone is replaced every two or three years and updates monthly. A car stays for ten. When native maps, voice assistants or app stores lag behind what people carry in their pocket, the gap shows up every single day. Buyers, who compare everything to their phone, will not be patient. A smarter answer may be a hybrid: own the vehicle and safety layers, open the experience layer. MG took the route with Hector, partnering with Microsoft, Adobe, Cisco, Gaana and others. Hyundai's Pleos Playground, an open platform that lets outside developers build in-car apps, points the same way.

India is not one market

India's SDV story only makes sense when you split it by vehicle type, because the economics are completely different. EVreporter's Q1 FY2026-27 data shows EV penetration of 9.2% in two-wheelers, 6.9% in passenger cars and 46.9% in passenger three-wheelers.

Figure 2. Three-wheelers lead India's EV shift. Source: EVreporter Q1 FY26-27 report, Vahan data

• Two-wheelers : EV share is still under 10%, but software-first brands such as Ather and Ola Electric already sold roughly 27,000 and 12,000 units in April 2026. For them the app, the updates and the dashboard are the product.

• Commercial vehicles : This is where connectivity has scaled fastest. Tata Motors says Fleet Edge has digitally connected five lakh commercial vehicles, and its commercial EVs, medium and heavy trucks and buses are Fleet Edge ready with embedded 4G SIMs. For a fleet owner, uptime is money, so a layer that tracks health and location pays for itself in ways a bigger touchscreen never will.

• Passenger cars : The idea is older than people assume. Maruti Suzuki has offered Suzuki Connect telematics since 2018 and extended it to Arena showrooms in 2021. Tata Motors now calls the Harrier.ev software its SDV foundation, built on acti.ev+ with more than 500 million lines of code and domain controllers, and Tata leads electric cars with over 36% share as of April 2026. Hyundai and BYD, discussed above, are the global stacks arriving from the other direction.

Three-wheelers: the SDV nobody is writing about

Three-wheelers are the surprise, and a uniquely Indian one, so they deserve more than a paragraph. Drivers cover long daily distances on thin margins, so running cost matters more than badges. Nobody sells these vehicles on screen size, yet they may generate the most useful real-world data in the country.

The timing matters too. The government's PM E-DRIVE target for L5 electric three-wheelers (three wheelers that have top speeds between 45 Km/h – 70 Km/h and can handle payload >  500 Kgs) was met ahead of schedule, with about 2.88 lakh vehicles supported, and that incentive window has now close. With the subsidy cushion gone, running cost and financing economics have to carry the segment, and this is exactly where software earns its keep. Our view of the priorities, in rough order:

• Battery health a lender can trust. A three-wheeler is a livelihood bought on credit. State of health, charge cycles and usage history decide both loan risk and resale value, and all of it comes from the battery management system.

• A cheap, standard connected layer. One telematics unit, a vehicle control unit and a BMS sharing a common data format, with over-the-air update capability. Think SDV-lite, built to a price.

• Energy flexibility. Software that handles swapping, charging and battery-as-a-service pricing, so the driver experience is similar to a ICE vehicle.

• Uptime. Predictive maintenance matters most for drivers who have no spare vehicle and lose income every hour the vehicle is down.

• Tuning by use case. Power modes and range management that adapt to payload, route and driver, delivered by update rather than a workshop visit.

Some players are already there. Euler Motors' HiLoad cargo three-wheeler ships with telematics for fleet tracking, battery monitoring and real-time charging, a 12.4 kWh pack and a certified range of 151 km, and Euler sells it alongside charging and maintenance as a full-stack service. There is also a policy tailwind. For L5 three-wheelers, e-rickshaws and e-carts, the revised phased manufacturing rules, enforced from May 2025, require traction battery packs, BMS, vehicle control units, DC-DC converters, onboard chargers and traction motors to be made in India to qualify for incentives. This creates a real opening for Indian-built control hardware and software.

How much of this is imported?

A caveat first. Nobody publishes a clean import share by vehicle segment, so what follows combines component-level estimates, policy documents and company claims. Treat the numbers as indicative.

The overall picture is consistent across sources. ICRA puts India's EV supply chain at only 30 to 40% localised. Chassis, body parts and some control units are made here, while battery cells, electronic control units and BMS parts are still largely sourced abroad, and cells, which can be 35 to 40% of vehicle cost in some segments, are entirely imported. IEEFA adds an important distinction. A component assembled in India can still depend heavily on imported chips and magnets, so high localisation does not automatically mean value created at home.

Figure 3. Estimated localisation of key EV components across segments, 2025. Battery pack range shown at its midpoint. Source: India EV & EV Components Market Report 2026-2032, as cited by Wright Research

Broken out by segment, the pattern looks like this:

The direction is encouraging. Organizations such as IEEFA and JMK Research find that several non-battery categories, including motors, power electronics and charging equipment, could reach 90 to 100% localisation by 2030 if announced capacity gets built. They also note that less than 10% of the INR 25,938 crore available under the PLI Auto scheme had been disbursed by early 2026. And the policy tends to measure process localisation, meaning where something is assembled, rather than value localisation. This gap is where India's software ambitions will be tested.

The Connected Scooters, different business

The EV discussion in India cannot be completed without a mention of Ather, Ola, Ultraviolette, in addition to the estimated 190 to 240 two wheeler EV manufacturers. We will talk about Ather as a place holder to discuss the positioning in the market.

Ather is often compared with Gogoro of Taiwan and NIU of China, since all three sell connected scooters. It is tempting to file Gogoro and NIU under fleet and rental, with Zypp as their Indian cousin. The numbers say otherwise. Both are mainly consumer businesses, with fleets a growing second leg. India's genuinely fleet-first player is Zypp, and the comparison is more revealing for that.

• Gogoro earns its keep in the network more than the scooter. In Q1 2026 it had 670,000 battery-swapping subscribers and a 69.4% share of Taiwan's electric two-wheeler market. Swapping revenue was $36.6 million for the quarter, and the company expects the network business to turn profitable on a non-IFRS basis in 2026, with hardware following in 2028. Fleets are a real part of the story: Gogoro said in 2023 that about 26% of Taiwan's electric-scooter food and parcel delivery riders used its swapping system, and it has run a green delivery programme with Uber Eats.

• NIU is a vehicle-and-app consumer play. It delivered about 434,000 scooters in Q2 2026, of which 402,202 were in China . China accounts for roughly 92% of its e-scooter revenue, and growth is helped by revenue from the Niu App, accessories and spare parts.

• Ather sits closer to NIU than to Gogoro. Its in-house AtherStack runs the touchscreen dashboard, navigation, ride modes, connected features and over-the-air updates, and the company earns both from selling scooters and from services such as Ather Stack. The software lives in the vehicle and the app.

• Zypp is the fleet-first mirror image. It reports over 22,000 IoT and AI-enabled EVs and is expanding with the Indofast swapping network. Every bike reports data on the bike, the rider, location, delivery count, battery status and maintenance into a proprietary fleet platform. Zypp also works with OEMs on made-in-India bikes where it controls the motor, battery and BMS technology. Gogoro led its $25 million Series B in 2023, so the two worlds are already connected.

The real contrast is less about scooters than about where the software sits: in the vehicle (Ather, NIU), in the energy network (Gogoro), or in the fleet operation (Zypp). India is likely to see a blend of all three, since swapping for delivery fleets is already under way.

Fleet or personal: the use case decides the stack

The same vehicle platform does a different job depending on who buys it, and it will shape how SDVs evolve here. A personal owner pays for experience and keeps the car for a decade. A fleet owner pays for uptime and cost per kilometre, and runs the asset hard. The table below is my framing of the difference.

Evidence from the market fits this split. Hyundai designed its Pleos OS partly to fold fleet management into the vehicle itself, removing the need for separate bolt on hardware and software. Tata's Fleet Edge shows what a fleet-first connected layer looks like at scale. Zypp's delivery-platform integrations and rider telemetry show the same logic on two wheels.

The implication is that India's SDV may split into two lineages. One is experience-led, covering premium cars and Ather-style scooters. The other is operations-led, covering three-wheelers, commercial vehicles and delivery two-wheelers. A modular update and telematics layer could serve both, and this is arguably what Hyundai is attempting. But the operations-led lineage probably has the larger and nearer-term base in India.

Key players at a glance

Where this leaves India

Candidly, with a big opportunity and a few uncomfortable gaps. India builds over 28 million vehicles a year and has deep software talent. But NITI Aayog also shows India holds just 3% of the traded auto components market, and imports about $1.4 billion of EV components against $0.1 billion exported, mostly from China. We are strong in code and thin in silicon and batteries.

Two-wheeler leaders have gone furthest in owning the stack, from BMS and controllers to the operating system, with cells and magnets the remaining dependence. Three-wheelers and commercial vehicles are being pushed by policy to localise drivetrain electronics, but chips and magnets stay imported for now. And Hyundai and BYD show the other end of the spectrum: complete platforms designed abroad that arrive with the vehicle.

India's SDV will not copy Munich, Silicon Valley, Seoul or Shenzhen. It will be affordable, data-rich and fleet-led, built first around three-wheelers, two-wheelers and commercial vehicles where the business case is obvious, and shaped by whether the vehicle is a livelihood or a lifestyle. Pair that with local compute, cleaner architectures and open experience layers, and India could export not just engineering hours but whole software platforms. The car is becoming an update and India has every reason to be the one writing it.

Disclaimer: We do not hold any positions in the companies mentioned above. This article is intended for informational purposes only