A modern vehicle now runs on around 650 million lines of code, and software is becoming the part of the car that defines its value. For automotive OEMs and suppliers, the 2026 question is no longer whether to build software capability but how to build, integrate, and validate it without the cost overruns and recalls that software defects already cause. This is a data view of where the spend is going and what is actually being built.
The numbers that frame the shift
According to Statista, the average vehicle is expected to run on roughly 650 million lines of code in 2025, after the per-vehicle figure roughly doubled from about 100 million to 200 million between 2015 and 2020. The McKinsey Center for Future Mobility projects the automotive software and electronics market reaching around 519 billion US dollars by 2035 at roughly 4.5 percent CAGR, with software development (including integration, verification, and validation) carrying a revenue potential near 83 billion US dollars by 2030.
Forecasts for the broader software-defined vehicle market diverge widely by analyst, which is itself a useful signal about uncertainty: estimates range from roughly 315 billion US dollars in 2025 (PS Market Research) to projections of 1.2 to 1.6 trillion US dollars by 2030 (MarketsandMarkets, BCC Research). The direction is consistent even where the magnitude is not: software is moving from a vehicle component to the vehicle’s defining system.
Why downtime data belongs in this conversation
Automotive software is built where automotive plants run, and the cost of getting it wrong is set by those plants. Siemens’ 2022 True Cost of Downtime study found the cost of an hour of downtime in automotive now tops 2 million US dollars, among the highest of any industry. Software that touches the line (manufacturing execution, supplier integration, and traceability) is therefore validated against a 2-million-dollar-per-hour failure cost, not a generic IT service level. That single figure explains why automotive software engineering is converging with an industrial-grade delivery discipline.
What OEMs and suppliers are actually building
- Connected-vehicle backends: telematics, over-the-air update services, and the cloud platforms that manage a fleet of software-defined cars after they leave the lot.
- Dealer and OEM platforms: configuration, ordering, warranty, and after-sales systems that tie the vehicle to the commercial operation.
- Supplier and plant software: traceability, quality, and integration systems for Tier 1 and Tier 2 suppliers feeding the line, where the 2-million-per-hour downtime cost lands.
- Legacy integration: connecting decades-old plant and ERP systems to the new software stack without replacing everything at once.
The last point is where most programs stall. The new software is built on modern stacks, but it has to interoperate with plant and ERP systems that may be 20 years old and undocumented.
The delivery lesson from outside automotive industry
The discipline that protects an automotive line is the same one that protects any production-critical system: preserve correct legacy behavior, migrate module by module, validate against recorded behavior, and run in parallel before switching.
On a 400,000-line legacy platform refactor in a regulated industry, this method cut manual refactoring effort by about 90 percent, delivered roughly 6 months faster than a conventional rewrite, and reached production with zero downtime. For automotive suppliers connecting old plant systems to new vehicle and OEM software, that zero-downtime, behavior-preserving method is a part of the playbook worth copying.
What this means for 2026 planning
- Treat software integration and validation, not feature development, as the cost and risk center. McKinsey’s own split puts large revenue potential specifically in development, integration, verification, and validation.
- Validate plant-facing software against the real downtime cost (2 million US dollars per hour for automotive), not a generic SLA.
- Never big-bang a cutover that touches the line. Parallel running is not optional at that downtime cost.
FAQ
How much software is in a modern car?
A: Around 650 million lines of code on average in 2025 per Statista, up from roughly 100 to 200 million between 2015 and 2020.
How big is the automotive software market?
A: McKinsey projects the automotive software and electronics market near 519 billion US dollars by 2035. Software-defined vehicle forecasts vary widely by analyst, from roughly 315 billion in 2025 to 1.2 to 1.6 trillion by 2030.
What is the costliest mistake in automotive software programs?
A: A big-bang cutover of plant-facing software. With automotive downtime topping 2 million US dollars per hour, an unvalidated switch on the line is the single most expensive failure mode.
Why does legacy integration matter so much in automotive?
A: New vehicle and OEM software has to interoperate with plant and ERP systems that are often decades old and undocumented. Replacing those safely, without stopping the line, is where most programs succeed or fail.
