Software Defined Vehicles
Interview with Heinz Schmitz, GreenRock
“Continuous pipelines have become business-critical”
As the shift toward software-defined mobility accelerates, OEMs face rising pressure across computing architectures, supply chains, and organizational models. Heinz Schmitz from GreenRock explains where the industry stands – and what companies must fix to keep pace.
Heinz Schmitz, founder and CEO of GreenRock, advises automotive companies on supply-chain resilience, organizational transformation, nearshoring, and SDV-related strategy. Before founding GreenRock, he held senior leadership roles across Tier-1 suppliers and international manufacturing organizations, managing global operations, restructuring programs, and technology initiatives in Europe and Asia. With long-standing experience in e-mobility, autonomous driving, and software-defined development, he is regarded as one of the industry’s most hands-on strategic advisors.
At the Automotive Computing Conference 2025, he engaged with OEMs, Tier-1s, and semiconductor leaders on the operational and structural challenges of the SDV transition. After the event, we spoke with him in depth about the state of automotive computing, the structural bottlenecks he sees at OEMs, and where demand for strategic support is currently highest.
ADT: Let’s start with a deliberately wide-ranging question: How do you assess the current state of automotive computing?
Schmitz: Automotive computing is entering a new era. Zonal architectures with centralized high-performance computers are becoming standard in premium segments, and OEMs are moving toward bespoke, AI-centric SoCs. Chiplet-based designs are emerging as the most cost-effective route to the multi-teraflop performance required by agentic AI and software-defined vehicles. The primary driver is the rise of agentic AI—systems that autonomously plan, prioritize, and execute tasks across vehicle domains. These capabilities accelerate development cycles, enable continuous feature delivery, and unlock functions that were impractical only a few years ago. The result is a fundamental rethink of both software and hardware architectures across the industry.
What aspects of this shift do you see as especially decisive?
Two hardware trends stand out. First, OEM-defined SoCs: manufacturers increasingly specify or develop their own SoCs with integrated AI engines and accelerators tuned to their use cases. This reflects the strategic value of AI and the need to optimize performance, power, and functional safety across domains. Second, chiplets and advanced packaging: growing compute demand pushes monolithic die sizes upward, reducing wafer yield and increasing cost. Chiplet architectures combine smaller, function-optimized dies into a single package using 2.5D and 3D packaging and high-bandwidth interconnects. The benefits include better yield economics, faster time-to-market through IP reuse, heterogeneous integration of CPU, GPU, NPU, I/O and memory, and greater supply-chain flexibility.
What does this mean for the way automotive systems will be built in the future?
The software and system implications are significant. Chiplets and modular SoCs promote standardized interfaces and middleware, improving portability and reuse across platforms. Functional safety and cybersecurity remain essential; hardware and software must be co-designed to ensure deterministic behavior, redundancy, and secure update mechanisms. Overall, we are seeing a structural transformation rather than incremental change. Agentic AI, SDVs, and modular hardware will drive the continued convergence of compute and software stacks. Broader adoption of chiplet ecosystems and tighter collaboration between OEMs, Tier-1 suppliers, and foundries will be necessary to balance performance, cost, and safety.
You advise companies in several strategic fields. What are the main supply-chain challenges your automotive clients bring to you today?
We frame supply-chain resilience around two primary risk scenarios. The first is single-source dependency. Certain components—particularly some processors—depend on a single supplier not only for the physical device but also for its proprietary development environment and toolchain. That combination creates a structural vulnerability that the industry clearly recognizes, and while new chip suppliers are emerging, it will take time to resolve. The second is geographic concentration risk. Heavy reliance on Southeast Asia, especially China, increases exposure to logistics, political, and capacity shocks. This concentration makes regional sourcing an attractive and pragmatic mitigation.
What do companies look for most when considering nearshoring options?
When advising on nearshoring within Europe, we classify Romania, Bulgaria, and Moldova as best-cost countries with competitive labor and operating costs, investor-friendly conditions, and a skilled technical workforce. These markets offer a balanced trade-off between cost and proximity to European OEMs and Tier-1 suppliers. Leveraging deep regional knowledge and an established network, we identify, audit, and qualify suppliers in these countries to meet automotive standards for quality, compliance, and scalability. Evaluated on a total cost-of-business basis—including logistics, ramp-up, and risk factors—nearshoring can deliver both improved resilience and tangible cost advantages. It is critical to recognize that these vehicles are not “smartphones on wheels.” Modern cars combine complex software and electronic systems with long-lead mechanical components such as body panels, injection-molded parts, and axles—each governed by very different development rhythms. Mechanical parts often require months from initial design to production-ready tooling, which in software terms can feel like centuries. That timing mismatch is the single biggest source of demand for our advisory services.
Where do you see the main organizational hurdles?
The two recurring issues we observe are asynchronous lifecycles and organizational legacy. Hardware and mechanical subsystems require long, deterministic lead times, while software development is rapid and iterative. Many OEMs and Tier-1 suppliers still operate with structures and governance models designed for vehicle programs from 10 to 15 years ago, which impede cross-domain delivery. We help clients align the long-lead hardware path with the fast software path through a combination of processes, organizational adjustments, and working tools. Every client is unique; there is no one-size-fits-all solution. We begin with a status-quo assessment and gap analysis and present our findings and proposals. A key success factor is agreeing with the client on a roadmap—with full management buy-in—followed by a pilot project and subsequent scaling and institutionalization.
As someone who has accompanied the Automotive Computing Conferences for many years, what were your key takeaways from this year’s ACC?
The influx of software and SoC players into automotive is reshaping who captures value; their technical and platform influence is shifting bargaining power and forcing suppliers and OEMs to rethink their offerings. Agentic AI and high-performance computing have moved from buzzwords to engineering practice. Validation and continuous pipelines have become business-critical, with CI/CT and scalable SIL/HIL strategies shortening iteration cycles and validating complex SDV architectures. Hardware trends taking hold include chiplets, zonal ECUs, and evolving supply strategies. My sixth time at ACC confirmed it as the premier conference for staying current and engaging with top experts.
