18 months. That is how long BYD takes from concept to market for a new car. Volkswagen announced in 2025 that it had compressed its own cycle to 40 months. The announcement was framed as a milestone. The gap has not narrowed: it has grown more visible, in the dimensions that matter most. But this is not about Germany vs. China, it’s about traditional systems engineering vs. Product Velocity.
The Product Velocity framework describes four principles that separate fast-moving product organizations from slow ones: Value Thinking, Architect for Flow, Shift Left, and Accelerate. Yuchao Luo, who worked inside both a Volkswagen joint venture and NIO, put those principles into concrete form in a recent interview. He did not come to celebrate China Speed. He came to explain it. The explanation is uncomfortable for anyone who wants to attribute the gap to cheap labor and state subsidies and leave it there.
Upcoming Webinar on Wednesday (May 27), 17:00 CET: How to Run a 30-Day Product Velocity Intervention >>
Value Thinking: Outcomes as Evidence
Value Thinking starts with a demanding commitment: define measurable outcomes and align all actors around them. In many organizations, velocity is constrained by ambiguity about what matters. Teams optimize for internal milestones while the actual value proposition stays implicit. China’s EV surge is, at one level, a government-defined value statement executed at industrial scale. Beijing decided in the early 2010s that electric vehicles were strategic. The country lacks domestic oil, and leading the EV transition meant owning the next cycle of mobility manufacturing. That intent became investment, which became a generation of engineers who treated EVs as the obvious future. BYD had placed the bet two decades before the wave arrived.
A Bosch China executive, quoted by Luo, stated the implication bluntly: a company that has sold a hundred units has no right to claim quality. The sample is too small to prove anything. Quality at scale is outcome, not process. Volume is the evidence of quality. This runs against the German reflex, which treats process rigor as the source of the quality claim. Compliance generates the certificate; the certificate certifies the outcome. At low volumes in safety-critical domains, this logic holds. At automotive scale, it can become a mechanism for producing certification artifacts without producing proof.
“Even though you set your process for a lot of compliance, for the very strict standards, you have no proof.”
Yuchao Luo, recounting a Bosch China executive
Institutional Rigor is Valuable
Germany’s institutional rigor is not worthless. ISO 26262 and ASPICE do catch real failure modes. German quality standards have saved lives in the field at volumes Chinese newcomers have not yet matched across all segments. That record is real and should not be discarded. The question is whether compliance is treated as evidence of quality or as a gate that must be passed before quality can be pursued. When compliance becomes the goal, organizations optimize for the audit rather than the user. The volume argument also does not apply everywhere: an aircraft manufacturer cannot sell millions of units. Within automotive, however, the challenge stands.
Architect for Flow: Platform vs. Legacy
Clear modularization allows teams to work semi-independently while the system stays coherent. Architecture that localizes change reduces coordination overhead and integration risk. China’s new EV players had a structural advantage here: no existing architecture to defend. Engineers from Huawei, Xiaomi, and consumer electronics entered car development without inheriting a network of 80 ECUs running on isolated communication buses. They built software-defined vehicle stacks from scratch. Low legacy burden is the phrase Luo uses. It describes freedom from documents, standards, and architectures accumulated over decades.
Platform thinking was not accidental. BYD’s vertical integration, from batteries to chips to software, gave it architectural control over its entire value chain. The Chinese government’s push for common charging standards and EV supply chain coordination created a shared platform at the industrial level. Companies competed on differentiation within that base.
Standardized Architectures Create Stability, But Only In a Stable World
German OEMs carry a different structure. Decades of supplier hierarchies, AUTOSAR compliance layers, and distributed ECU networks reflect decisions that were correct when they were made. AUTOSAR deserves some credit: it created a shared middleware standard across competing companies and supplier tiers, a coordination achievement that required years of industry alignment. But AUTOSAR was designed for a world of fixed hardware and infrequent software updates. The software-defined vehicle inverts those assumptions, and legacy architectures resist inversion. The coordination capability that produced AUTOSAR now needs to produce something structurally different. That requires acknowledging the legacy as a constraint before it can be managed as one.
Shift Left: The Order of Questions
Shift Left means moving learning earlier. Uncertainty discovered late is economically expensive. Uncertainty removed early is cheap. The cost of change does not drop linearly as development progresses. It escalates. China’s new EV players shifted their feedback model entirely. Direct user communities replaced dealerships. OTA updates within quarters replaced end-of-program validation cycles. The mechanism was not informal: NIO and XPeng both built the infrastructure to hear a user and respond within a product cycle.
The Chinese government’s OTA regulation is the clearest example of Shift Left applied at a regulatory level. When critics raised the risk that OEMs might use silent OTA updates to patch safety problems without disclosure, the government did not ban OTA. It required reporting. An OEM that wants to push an update must declare its content to the relevant authority before deployment. The authority can take months to clear a release. The iteration channel stays open. The response to “fast iteration is dangerous” was to make iteration visible and accountable, not to close it.
“If at the beginning of one innovation you directly talk about compliance, the government is normally slower than the industry. At that time there is no compliance for this new innovation.”
Yuchao Luo
Germany Has the Tools, But Must Use Them Differently!
Germany’s regulatory reflex tends in the opposite direction. Luo describes European engineers, on hearing that AI can produce requirements at 70 to 80 percent accuracy, asking first how correctness can be proven. Chinese engineers ask the same question, but after using the output. They take the 70 to 80 percent, apply human review, and correct what the model got wrong. Humans also make mistakes. The industry already runs reviews and processes to catch human error. Holding AI to a standard of zero error that humans were never held to is not caution. The difference between the two cultures is not awareness of risk. It is which question comes first.
German automotive already has the tools. HIL test clusters, automotive CI/CD tooling, ASPICE process requirements for supplier qualification: genuine Shift Left instruments, in use. The tools are not the problem. The sequence is. Organizations apply compliance gates before they understand the problem, rather than after the design has stabilized. That order undercuts the tools and prevents them from doing what they were built to do: accelerate feedback, not defer it.
Accelerate: Closing the Loop
Accelerate requires that release mark the beginning of a learning phase, not the end of a development phase. Operational feedback must flow back into architecture and business decisions. Without this, the other three principles produce a well-structured product that stops evolving the moment it ships.
China’s EV near-collapse of 2020 and 2021 is instructive. NIO’s share price fell toward one dollar. XPeng faced similar pressure. The companies that survived did so by staying close enough to their users to hear what was failing and iterate in response. The feedback loop was not smooth: supply chains, COVID, and individual incidents all hit simultaneously. The organizational reflex, learned from mobile-phone development, was to keep the loop turning rather than stop and replan.
Germany Must Keep the Learning Relationship Open
Volkswagen’s response in China is more substantive than European commentary usually credits. Investment has shifted from manufacturing plants toward R&D talent. Joint ventures with Horizon Robotics and XPeng are attempts to learn from competitors rather than dismiss them. That choice, to stay in China and invest through geopolitical friction rather than exit, keeps the learning relationship open. Luo credits this not as a transformation complete but as evidence that a legacy OEM can adapt when competitive pressure is sufficient and leadership accepts what that pressure is telling them.
Germany’s strength in the Accelerate dimension is institutional durability. Companies that leave a market under pressure lose the operational feedback loop entirely. Staying, even when staying is difficult, keeps the channel to the field open. That matters precisely because Accelerate depends on the relationship between development and operation remaining live. An OEM that withdraws from its fastest-iterating market loses both the revenue and the learning. That second loss is the more serious one.
Conclusion
The four principles are not a Chinese invention. They describe what happens when any organization is structured around learning speed and value flow. China’s EV industry applied them not by design but by circumstance: new entrants with no legacy, a government with a clear strategic objective, engineers with a mobile-phone iteration reflex, and regulators who learned to make iteration accountable rather than impossible.
Germany has strengths that the China-speed narrative consistently underplays. Its engineering rigor catches failure modes that would be expensive at scale. Institutional stability in Germany supports long-term supplier relationships and market commitments. Its engineering depth produces systems that have been tested against conditions most startups have not faced. These are real advantages, and they are why German automotive is not finished.
They are insufficient when the competitive question has shifted from “is it good enough?” to “can you get there before the customer stops waiting?” Luo’s clearest observation is about the order of questions. Ask whether something is compliant before it exists, and the innovation never reaches the point where compliance would be meaningful. Ask whether it works and can be proven at scale, and compliance becomes a constraint to integrate rather than a gate to hide behind. The European instinct to raise concerns is an asset in a mature product line. It is a liability at the start of a new one. Knowing which situation you are in is the whole job.
English
German