By the end of 2025, many ideas that recently sounded experimental have moved into production: software-defined vehicles (SDVs) are no longer just roadmaps, zonal and domain architectures are replacing ECU sprawl, Android Automotive OS is showing up in real cars, and artificial intelligence (AI) is starting to influence not only driver assistance but also the everyday in-cabin experience. The automotive industry may still move carefully, but this is where things are clearly heading.
We spoke about this with Maxim Leykin, PhD, Chief Technology Officer at Bamboo Apps. Maxim has over 20 years of experience in software development and has been involved in projects for leading automotive and industrial technology companies.
This conversation looks at how automotive software is being built in 2026 – what will become the new “default”, and what will still be painful and fragmented.
Maxim, how would you describe the biggest shifts in automotive software development you’ve seen in the last 12–18 months?
Such shifts are partly caused by changes in the automotive industry itself. For example, more and more attention is currently being paid to autonomous vehicles, which implies strict requirements for the reliability and security of automotive applications. In turn, this leads to improved standardisation, e.g., AUTOSAR Adaptive adoption.
AUTOSAR Adaptive is a modern software standard for powerful vehicle computers, designed for complex tasks like autonomous driving, connectivity, and over-the-air updates, differing from Classic AUTOSAR by using POSIX OS, C++, and service-orientated architecture for dynamic, high-performance systems, while Classic focuses on traditional, hard real-time embedded functions.
Another important shift is related to the SDV architecture. Previously, OEM manufacturers mainly discussed SDVs – now many are rolling them out in production vehicles. E/E (electrical/electronic) architectures are moving from function-specific ECUs (electronic control units) to domain controllers and zonal compute. This enables OTA (over-the-air) updates, real-time reconfiguration, and extends the life cycle of the vehicle.
According to a Deloitte study, the shift toward software-defined vehicles is already underway. More than 90% of surveyed OEM executives report active adoption of SDV strategies. Deloitte itself projects that by 2030 around 81% of OEM vehicle fleets will be software-defined.
Which technologies or trends do you think will redefine automotive software in 2026?
The automotive industry is quite sluggish, so I would not say that new technologies and trends will radically redefine it. Instead, it will use technologies that are already trending in other areas. Unsurprisingly, artificial intelligence is the key technology, and it will increasingly influence both operational software and user experience, changing the way drivers and passengers interact with vehicles.
Moreover, artificial intelligence is expanding into broader areas, for instance:
- Real-time sensor fusion and decisioning for safety systems, e.g., automatic emergency braking (AEB) activation if a pedestrian is detected and driver reaction time is insufficient.
- Predictive maintenance and vehicle self-optimisation, such as automatically suggesting a charging stop based on your real-time driving style and current battery health, rather than just the remaining mileage.
- Natural language and context-aware assistants inside the cabin, e.g., a driver can say, “I’m feeling stressed, find me a scenic route home and put on something relaxing,” and the car will adjust the navigation, ambient lighting, and massage seats simultaneously.
From your perspective, how do you see automotive software stacks evolving in 2026 compared to today?
I expect that in 2026, software stacks for the automotive industry will become less fragmented since, in my opinion, high fragmentation is one of the most serious problems that is delaying progress nowadays. I also expect automotive software to become more open and less proprietary. Finally, I anticipate that Android Automotive OS will become dominant in the industry in 2026 or at least in the next few years.
Read our overview of the ways to elevate in-car experiences with Android Automotive apps here.
In practice, at Bamboo Apps we are actively working with such technologies as Android Automotive OS, Linux FreeRTOS, Kanzi, and BlackBerry QNX.
What technological trends are most responsible for the recent shift in how vehicle software is built?
The way we think about vehicle software today is fundamentally different from even just a few years ago. For me, the big shift is that cars now are connected edge platforms. We’re seeing this in several ways:
- Vehicle-to-Everything (V2X) communications are helping cars talk to traffic infrastructure and other vehicles, improving coordination and safety.
- Car-to-Cloud (C2C) pipelines let us stream data, run analytics, and push remote updates, so software can evolve even after the car leaves the factory.
- Edge AI is becoming crucial for making decisions in real time without waiting on a cloud connection.
All of this means the software boundary isn’t just the vehicle anymore. It’s the entire mobility ecosystem.
The industry at large is backing this up with real commitments: at CES 2026, AWS announced expanded collaborations with companies like Aumovio to support AI‑driven development of autonomous vehicles, leveraging advanced cloud AI capabilities such as Amazon Bedrock and generative/agentic AI to speed up testing, validation, and simulation workflows.
Rather than manually sifting through millions of miles of sensor and driving data, engineers can now query vast scenario datasets in natural language and automatically surface rare “edge cases”. For example, pedestrians entering a roadway at night in rain, which are critical for validating Level 4 autonomy. This work is being applied to industrialise and scale the Aurora Driver for autonomous trucks, with AWS processing enormous volumes of sensor and simulation data to help satisfy thousands of safety requirements and millions of virtual test cases before deployment.
With so many new technologies entering vehicle software, what technical skills will be non-negotiable for automotive developers in 2026? Are there also emerging soft skills or cross-disciplinary abilities that will become crucial?
In my opinion, Rust is now non-negotiable for new safety-critical modules. Developers must understand “Memory Safety” and “Ownership” models to eliminate bugs like use-after-free at compile time.
Alongside language proficiency, knowledge of system-level orchestration is crucial. Proficiency in Hypervisors (like QNX) and Automotive-grade Docker/Kubernetes is required to run mixed-criticality apps (e.g., ensuring a Spotify crash doesn’t break data on a speedometer).
Cybersecurity is equally important. Developers must be fluent in Threat Analysis and Risk Assessment (TARA) and understand how to implement HSMs (Hardware Security Modules) for secure communication.
In addition, it is also mandatory to know how to integrate LLMs and Vision AI into the car’s middleware. This includes MLOps for cars – managing the lifecycle of a model as it receives OTA updates.
As for non-programming skills, I believe that in 2026 automotive developers should play the role of “translators” who can bridge the gap between various engineering worlds, which includes fluency in various fields (Mechanical <-> Software) and an understanding of Vehicle Dynamics (how the car physically reacts to torque) to ensure the software feels natural and safe to a human driver.
What does an ideal automotive software development team look like in 2026?
There is no perfect team, just as there is no single optimal one for all cases.
In fact, team composition is determined depending on the project’s nature and requirements. To develop software for the automotive industry, it is usually important to involve system architects familiar with automotive E/E systems, software and security compliance specialists, and user experience designers.
Could you share a story about a lack of skills affecting project implementation or innovation?
A few years ago, we started developing our internal R&D project, “In-Vehicle Onboarding”, which was supposed to have Android, iOS, and AAOS applications. We successfully published the Android and iOS versions in their respective stores. However, the AAOS app was initially rejected by Google after its platform review.
Back then, we spent a significant amount of time and went through several unsuccessful attempts to pass the Google review for automotive applications, receiving different rejection reasons each time. The main challenge was that Google’s requirements for Android Automotive app submission were somewhat controversial, especially when specifying and mixing app categories. Our lack of prior experience with this process only added to the challenge.
Nevertheless, after careful analysis and investigation, we were eventually able to meet all requirements and publish the app successfully.
Are you keeping pace with how automotive software is evolving?
Get in touch with Bamboo Apps and explore how to move from trends to production.
Conclusion: what to take away from this conversation
Looking at the last few years, automotive software hasn’t changed because of one sudden breakthrough. And the role of software inside the car is expanding. Vehicles are becoming edge platforms connected to cloud services and infrastructure, while AI is moving from research demos into practical use cases. This shift doesn’t radically reinvent the industry overnight, but it clearly raises the baseline for what automotive software is expected to do.
For developers and teams, 2026 looks less about chasing new frameworks and more about mastering system-level thinking. Memory safety, security, mixed-criticality orchestration, cloud and edge integration, and the ability to bridge software with vehicle behaviour are still essential.
At Bamboo Apps, we work with these realities every day. If you’re navigating the shift toward software-defined vehicles or exploring what your automotive stack should look like next, we’re always open to continuing the conversation. Feel free to contact us!
