Software Defined Vehicles
Longstanding engineering practices suddenly feel insufficient
The New Realities of Automotive Software Development
There are moments in an industry’s evolution when longstanding engineering practices suddenly feel insufficient, even to those who helped define them. Automotive software development is experiencing such a moment today.
As vehicle architectures shift toward high-performance compute platforms, continuous software deployment, and complex cross-domain interactions, developers are encountering problems that cannot be solved with the tools of the past. BMW’s recent decision to adopt Percepio’s observability technology is a visible example of how OEMs are modernizing their toolchains to address these new realities.
The findings from QNX’s Under the Hood: SDV Developer Report (October 2025) provide important context for understanding why this shift is accelerating. The report gives voice to thousands of automotive developers working inside global OEMs and Tier-1s. Their message is consistent and candid: debugging, testing, and integration complexity are among the biggest obstacles facing SDV programs today.
The Bottleneck Developers Talk About Most
One of the report’s central insights is that automotive developers feel constrained by the difficulty of diagnosing software issues across increasingly complex runtime environments. More than a third identified debugging and testing as a major barrier to SDV progress. The same proportion highlighted integration complexity, while others pointed to long development cycles and insufficient tools in their daily workflow.
Advertisement
That combination of challenges should sound familiar to anyone working at the intersection of embedded and distributed software systems. Traditional debugging practices simply do not scale to multi-core compute clusters, mixed-criticality workloads, multi-threaded applications, and asynchronous communication patterns.
Developers want tools that can illuminate behavior across these boundaries. They want to stop guessing and start seeing.
The Shift Toward Developer-Centric Environments
Another striking finding in the QNX report is that only 30 % of developers describe their development environment as excellent. In most industries, such an admission would raise alarms. In automotive, it has become a quiet but universal frustration – one that is now undermining schedules and quality targets.
BMW’s investment in stronger observability reflects a broader shift toward prioritizing developer experience. The industry has realized that development environments are no longer a peripheral concern; they are integral to SDV success. If developers lack insight into runtime behavior, they spend more time searching for faults than creating value.
This is where modern observability transforms the engineering process. Instead of manually instrumenting code, adding diagnostic logs, and attempting to recreate timing-sensitive issues in the lab, developers gain a continuous view into how software behaves – on real hardware, in real workloads, under real conditions.
Recalls, Risk, and the SDV Lifecycle
The QNX report also reveals how deeply software recalls have shaped developer psychology. Nearly 60% of developers say recalls have significantly adjusted their approach to software development, a number that has grown steadily in recent years.
Advertisement
The message behind this statistic is clear: developers are acutely aware that small defects can have outsized impact in a software-defined vehicle. Timing jitter in a sensor fusion module, a missed edge case in an over-the-air update flow, or an unexpected inter-process dependency can escalate into system-wide instability.
Observability is one of the few technologies that helps teams detect such conditions early – and understand them reliably. For OEMs, this translates into faster triage, fewer field failures, and greater resilience in the face of regulatory frameworks that now extend across the entire software lifecycle.
AI Will Only Increase the Need for Better Data
The QNX report highlights another notable theme: developers expect AI to play a major role in automotive software development within three to five years. Many foresee AI-driven code analysis, automated bug detection, and intelligent test generation becoming core tools in automotive engineering.
But AI thrives on high-quality data, and runtime trace data is among the most valuable datasets available for training models that detect anomalies, regressions, and inconsistent execution patterns. Without reliable observability, AI-assisted development risks becoming little more than an idea.
In other words, as AI enters the SDV toolchain, observability becomes even more indispensable.
What BMW’s Decision Tells Us
BMW did not need a report to understand these pressures; like other OEMs, it feels them directly within its own development organization. But the alignment between BMW’s strategic choice and the QNX report’s findings is noteworthy.
BMW’s adoption of Percepio’s technology speaks to a clear recognition: runtime visibility is now a core requirement for SDV development, not an optional enhancement. It reflects a shift toward proactive diagnostics rather than reactive debugging, toward system-level understanding rather than module-level assumptions.
This shift is happening across the industry, but BMW’s move provides a concrete example of an OEM willing to modernize its toolchain in response.
The Road Ahead
As vehicle architectures continue to centralize and developers assume responsibility for larger and more complex software stacks, observability will become a foundational part of the engineering ecosystem. It will support everything from performance tuning to safety validation, cybersecurity monitoring to OTA update verification.
The QNX report shows that developers want this evolution. BMW’s decision illustrates that OEMs are enabling it. And as the SDV era accelerates, runtime insight will become one of the essential capabilities that separates successful development programs from the rest.
Percepio’s role in this transformation is to equip developers with the visibility they need to understand their software, validate their assumptions, and build the next generation of automotive systems with confidence.
The industry is ready for this change – and BMW is among the leaders demonstrating what it looks like in practice.
