Dimitar Dimitrov is the founder and Managing Partner at Accedia, a leading European AI & Custom Software Development Company.getty​Over the past year, I have spoken with technology leaders at automotive manufacturers and suppliers, and one pattern keeps repeating. They have grown their software teams faster than ever, yet they keep missing their launch dates.The software-defined vehicle has become the industry's central competitive challenge. In these vehicles, the features and the driving experience are delivered and updated through software rather than fixed hardware, which puts software at the center of the product. For 45% of the manufacturers and suppliers surveyed by IoT Analytics in 2026, this shift is now their single most important strategic priority, ahead of driver-assistance systems and electric vehicles. They are all pursuing it at the same time, competing for the same narrow pool of engineers who can build the software that differentiates a vehicle.In my experience, the size of a manufacturer's software team tells you little about whether it will deliver. What matters more is what its best engineers actually work on. The manufacturers moving fastest are deliberate about that. They make three decisions early, about what to build in-house and what to hand to a partner, that most teams put off until a program is already in development.Why Engineering Talent Has Become The Binding ConstraintThe engineers who can build a vehicle's core software, its operating system and its safety-critical controls are rare in every market, and no hiring plan closes that gap on the schedule a board has in mind.Yet in most programs, these are the very people absorbed by work that adds nothing a buyer would notice. They run cloud environments, maintain over-the-air updates, staff internal support and handle compliance. All of it is necessary, but none of it decides whether a customer picks one vehicle over another.One head of vehicle software put it to me plainly. His most experienced engineers spend their days keeping existing systems running, while the features meant to set the next model apart keep slipping. That gap shows up a model year later, when the capability that was supposed to justify the price is still in testing, and a competitor has already delivered its own.Three Decisions To Make Before The Project BeginsMost manufacturers back into these decisions while a program is already being staffed, and the default answers are expensive. Each one can be made on purpose instead, and each decides where your strongest engineers end up.1. Identifying The Software That Differentiates Your VehiclesStart by separating the software that differentiates your vehicles from the software that simply has to work. The test we give clients is simple. If every vehicle in the segment eventually has a capability, a partner can build it. If it shapes how your vehicles drive or how they improve after the sale, it belongs to your own engineers, the ones who are hardest to hire.Most programs never draw this line and treat every piece of software as equally important. That is how senior engineers end up scattered across work a partner could handle, while the features customers will notice wait their turn.The capability test: Can you name, today, the few capabilities that only your own engineers should build?2. Delegating The Supporting Layers To A PartnerMany manufacturers assume that controlling their software means building all of it themselves. The instinct is understandable, since software is increasingly where the recurring revenue sits. On the evidence, though, trying to build everything in-house is what slows delivery down.S&P Global Mobility reported in January 2026 that the fastest manufacturers, led by Chinese automakers, now go from concept to production in roughly 18 to 24 months, against an industry average closer to five to seven years. The difference is focus, since the quickest manufacturers are deliberate about how much they take on themselves.Ford shows the cost of the opposite approach. It spent years and billions of dollars building its own next-generation vehicle software and electrical architecture, then canceled the program in 2025 when the costs and delays ran past the plan, according to Reuters. The ambition was understandable, but the project's scope appears to have consumed significant time and engineering resources before it could deliver on its intended software advantages.​The make-or-buy test: For every layer you plan to own, can you explain what owning it gives you that a capable partner could not?3. Sequencing The Ownership Decisions Before Development StartsThe order of these decisions matters as much as the decisions. Deloitte puts the software-defined vehicle market at $400 billion to $600 billion by 2030, so misplacing your strongest engineers is an expensive mistake to leave for later. Starting development before ownership is settled costs you twice over. Teams rebuild the same component because the ownership question comes back the moment two teams touch the same code.​One example comes from our long-running engagement with a large European automotive group. The project focused on building and operating the high-availability cloud platform and automation supporting the company's digital vehicle services, a complex effort involving cloud architecture, security and large-scale automation. The platform has maintained more than 99.9% availability, while release cycles have been reduced from weeks to days. More importantly, clearly defined ownership of the underlying platform allowed the client's engineering teams to focus on developing customer-facing capabilities instead of repeatedly addressing foundational infrastructure issues.​The sequencing check: Have you settled who owns each layer before development starts, or will that get decided under pressure later?ConclusionThe manufacturers pulling ahead share a habit. They decide early what their own engineers must own, hand the rest to a partner and hold that line as the program grows.Adding headcount is the easy move, but the harder and more valuable one is choosing where that talent goes, and making that choice before the first line of code is written.Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?