Augustin Friedel on What OEMs Need to Get Right to Succeed in SDV and AV Transformation

In this interview with Focal Point Positioning, Augustin Friedel—SDV expert and Associated Partner at MHP—shares a candid view of what it takes for incumbent OEMs to make the leap to software-led operating models, how to think about the path to a full software-defined vehicle (SDV), and what it will take to build customer trust as autonomy scales.

FocalPoint: You guide OEMs through SDV transformation roadmaps. What is the biggest organisational barrier you see when large, legacy companies move towards software-led models?

Augustin Friedel: Willingness and ability to change is the number one barrier, and it shows up in subtle but systematic ways. Even when top decision makers signal a clear intent to move toward software-led practices, other streams in the organisation often resist or quietly avoid change. Many legacy firms resemble ‘ice‑cream sandwich’ organisations: enthusiastic vision at the top and strong execution energy on the front line, but a frozen middle that preserves the old structures, incentives, and governance. That frozen layer makes it nearly impossible to realign roadmaps, budgets, and accountability from component-centered projects to product platforms and services. Breaking through requires a mandate that is both structural and cultural: product-based P&L ownership, multi-disciplinary teams with true end‑to‑end accountability, and operating cadences that reward learning speed over plan conformance.

What is one mistake you would advise automotive leaders to avoid?

Do not wait. The cost of doing nothing is now higher than the cost of imperfect action. Standing still erodes competitiveness on multiple fronts: customer experience, cycle time, and the ability to attract and retain software talent. Leaders should not look left and right to see who moves first; they should take the lead for the area they own and demonstrate the behaviours they expect from others. Lead by example: ship improvements in shorter cycles, remove handovers, and make decisions that favor reusable platforms over bespoke one‑offs. When leaders model change, organisations follow.

Cost pressure is rising while safety requirements are increasing. Where can software meaningfully replace hardware complexity without compromising safety?

The most reliable savings come from consolidating hardware into fewer, more capable compute domains and letting software carry the variation. Instead of proliferating ECUs and wiring for every feature variant, OEMs can adopt a common hardware baseline and enable or refine functions via software, feature flags, and calibrated models. Sensor fusion and perception improvements can substitute for some single‑purpose hardware by extracting more information from existing sensors, while rigorous safety concepts—fault containment regions, health monitoring, and graceful degradation—ensure we do not trade away safety. Software‑defined HMI can reduce physical switches without reducing usability, and predictive diagnostics can cut warranty spend by identifying issues before they cascade. The rule is simple: remove hardware only where you can prove, through evidence and safety cases, that software plus system design preserves or improves the residual risk profile.

As vehicles move towards Level 3 and beyond, how should OEMs think about confidence in their sensor and positioning stack? What role does redundancy play in building that confidence?

Location intelligence is now a core element of advanced driver assistance and autonomy. High‑integrity positioning—using highly reliable GNSS and fusing it with inertial sensors, wheel odometry, and perception landmarks—raises availability and resilience, especially in urban canyons and adverse weather. We see different philosophies across ADAS levels: some suppliers bet on AI‑ and compute‑heavy perception to reduce sensor counts, while others maintain a richer, heterogeneous stack to maximize observability.

For Level 3 and above, redundancy is a design requirement because the machine is, at times, in charge. Redundancy spans sensing (diverse modalities), compute (fail‑operational paths), power, and actuation. But naive redundancy drives cost and weight, and current brute‑force approaches are not sustainable. The way forward is evidence‑based optimization: use scenario‑driven safety cases, quantitative performance envelopes, and continuous fleet telemetry to determine where diversity is essential and where advances in software, perception, and self‑diagnostics deliver equivalent or better confidence at lower cost.

What does it actually take to go from a hybrid SDV to a full SDV? Where do OEMs consistently underestimate the difficulty?

Reaching a full SDV is not a technology upgrade; it is a different operating model. Of course the technical foundations matter: centralized compute, a service‑oriented software architecture, a robust middleware and API layer, safe over‑the‑air (OTA) pipelines, and production‑grade telemetry. But the hard part is organisational: shifting from project funding to product lifecycle funding, from functional silos to platform teams, from milestone acceptance to continuous verification and validation. OEMs often underestimate the difficulty of aligning compliance, safety engineering, and cybersecurity with continuous delivery. They also underestimate what it takes to operate vehicles as connected products at scale. Think beyond the product in the factory; think about the product as a continuously evolving service in the field.

Customer trust becomes essential as autonomy advances. How can OEMs and ecosystem players build and sustain that trust—and does this differ across regions?

Trust is earned through clarity, consistency, and responsiveness. First, explain the technology in plain language—what the system can do, where it works, and where it does not. Second, set clear guardrails by defining and communicating the operational design domain, driver responsibilities, and handover behaviors. Third, follow the rules: adhere to local regulations and standards and be transparent about data usage. Fourth, fix fast—operate with an incident response mindset, ship corrective updates quickly, and close the loop with customers. Finally, build a ‘story machine’: a steady flow of real examples, metrics, and user stories that show improvements over time. Regional nuances matter. Europe typically expects stronger privacy protections and rigorous regulatory alignment; the U.S. values clear liability and practical benefits; China emphasizes rapid scaling, tight ecosystem integration, and regulatory compliance framed by national priorities. The principles are universal, but the narrative, evidence, and engagement model must be localized.

What is the role of partnerships in driving innovation in the SDV and ADAS space today?

Delivering SDVs and increasingly automated driving vehicles is not possible without partnerships. The stack is simply too broad—from silicon and toolchains to perception, mapping, cloud, and user experience—for any single company to excel end‑to‑end. Partnerships unlock speed, share cost, and enable hyper‑localization to meet regional customer expectations. At MHP, we recommend a three‑phase approach.

• First, enlightenment: align on outcomes, define the operating model and risk sharing, and establish the common language and architecture principles.
• Second, execution: stand up joint product teams with integrated roadmaps, shared telemetry, and contractual frameworks that reward delivered capability, not just effort.
• Third, evaluation: instrument the partnership with clear KPIs and health metrics, run regular business and technical reviews, and be willing to pivot or rebalance scope as evidence emerges. Strong governance and a culture of transparency make partnerships a force multiplier rather than a coordination tax.

In short, success in SDV and autonomy will favour the OEMs that change how they operate as much as what they build: fewer hardware variants, stronger software and safety disciplines, faster feedback loops, and ecosystems that compound learning rather than fragment it.

Augustin Friedel, Associated Partner at MHP, is one of the leading industry experts, particularly known for his expertise related to SDV strategies, autonomous driving, and mobility services. He reaches more than 59,000 followers on LinkedIn and his focus is on guiding OEMs, Tier 1 and Tier 2 suppliers through the transformation towards Software-Defined and AI-Defined Vehicles.