When Germany Seized the Future: Lessons from CAD for the AI Era
In his November 2025 interview with Dwarkesh Patel, Ilya Sutskever articulates the biggest dissonance plaguing today's AI community: while benchmark results for new LLMs get better and better, the high-value industry applications are mostly missing. He expresses genuine confusion — the models "look so good on benchmarks but fail so badly in practical applications and have yet to impact the economy in any meaningful way."
Talk to AI experts in enterprises and government, and you hear frustration between the lines. The lack of transformative applications is not through a lack of trying: enterprise AI spending has surged from less than $2 billion in 2023 to about $37 billion in 2025. Data, models, proof-of-concept implementations see billions invested, because we all agree: this industrial revolution will disrupt every industry - now is the time to build empires. And lose them.
A whopping 74% of companies have yet to see tangible value from their AI initiatives. 42% of companies abandoned most of their AI initiatives in 2025—more than double the rate from just a year earlier. The gap between AI promise and AI reality has never been wider.
There was rarely a time in technology history producing more fertile ground for bold disruptive strategies. But also: this is a time where the most costly mistakes will happen — and nobody can afford to be asleep at the wheel.
A Crisis with a Precedent
While AI turns all of knowledge work upside down, we are also in the middle of a major shift in automotive and manufacturing. The move towards EVs, plus a new geopolitical world order targeting sovereignty and manufacturing with massive industrial policy interventions, is changing all the rules.
German and European players are in crisis. In 2025, traditional Car manufacturers' markets are collapsing, job cuts are looming. German market share in China has collapsed to just about 15% - down from 25% just a few years ago, while domestic brands now control about 70% of the Chinese market. The old playbooks have stopped working.
Can Germany arise from this turbulence among the winners? The pressure is immense: a demographic collapse, a geopolitical shift, changing global market preferences, and the overwhelming might of trillion-dollar tech companies.
Although today's circumstances are unique, there is a period from recent manufacturing history that shares striking similarities—and had Germany emerge on top of the world. It's worth looking back to the last game-changing technical shift in manufacturing.
The CAD Revolution: A Story of Second-Mover Triumph
For roughly 30 years, German engineering was on top of the world. A significant part of all groundbreaking automotive innovations came out of Germany, and the German luxury brands came to command around 80% of the global premium automotive market. The interesting thing: this strength was a direct result of a bold foundational transformation that didn't originate in Germany, nor was the technical IP owned by German companies.
But the new capabilities found in the then second-largest economy a uniquely fertile soil — and decision makers who seized a once-in-a-lifetime opportunity.
The ingredient of Germany's automotive dominance from the 1980s onward came with a paradigm shift based on general-purpose technology: the rise of computing and computer-aided design, simulation, and manufacturing planning with sophisticated computer programs instead of relying on whiteboards, clay, and pencils.
General Motors and IBM built the first automotive interactive graphics system, DAC-1, in 1964 — but didn't put massive additional effort into the technology. Then the French innovated crucial methods: Pierre Bézier at Renault developed UNISURF by 1972, creating mathematical curves that would become foundational to all surface modeling. Paul de Casteljau at Citroën developed parallel approaches as early as 1958 — but Citroën kept his work secret until 1974, limiting its impact.
In the early 1970s, Toyota's TINCA and Mercedes-Benz's SYRCO took sophistication a step forward, but again with limited broader impact.
The crucial pillar of German dominance was arguably created through a joint venture at the end of the 1970s: Volkswagen commissioned ICEM Systems GmbH to build a 3D surface modeling system called VW-Surf. This system would eventually become the industry standard for automotive surface design. By the 1990s, many car manufacturers switched from their proprietary developments to ICEM Surf for creating the "visible surfaces" of automobiles.
Meanwhile, French aircraft manufacturer Dassault had been developing CATIA (Computer-Aided Three-dimensional Interactive Application) since 1977, initially for designing the Mirage fighter jet. When CATIA arrived in German automotive plants in the mid-1980s—with Francis Bernard, CATIA's inventor, personally demonstrating the system to Mercedes, BMW, and Honda engineering divisions — it didn't create German precision. It amplified an existing system that was uniquely positioned to exploit it.
The Amplification Effect
The introduction of CAD was not merely a new tool for the worker; it was a new nervous system for the industry. By moving the difficult, error-prone work of defining geometry from the physical world to the digital realm, German automakers leveraged their inherent engineering strengths at an unprecedented scale.
The results were dramatic. By the 1990s, the manufacturing sector's productivity growth accelerated from 2.6% per year in the 1980s to 4% per year — and the most computer-intensive manufacturing sectors showed labor productivity growth of 5.7% annually, compared to just 2.6% in less digitized sectors. Modern implementations show even more striking numbers: one large German automotive manufacturer achieved a 93% reduction in design review time — from two weeks to two hours.
But the technology alone wasn't the differentiator. What made Germany win?
The Secret Ingredient: Fertile Soil
Germany's rise wasn't about being first with the technology. America invented CAD. France pioneered the mathematical foundations. But Germany perfected the implementation.
This pattern has a name: second-mover advantage. Research shows that in technology adoption under uncertainty, followers often accrue advantages from better information about the value and costs of technology, lower risk, and the ability to learn from pioneers' mistakes. In one landmark study, first-movers were more successful than late movers in only 15 of 50 product categories.
What Germany had was something money couldn't buy: a culture of precision, reliability, and efficiency deeply ingrained in the national psyche, traceable to artisanal traditions going back centuries. A dual education system combining classroom theory with hands-on practice, embedded deeply within society. And a Mittelstand of 1,573 "Hidden Champions"—market-leading, innovation-driven small and medium-sized enterprises, many holding 70-90% world market share in their niches.
CAD didn't create German engineering excellence. It gave German engineering excellence the tools to scale beyond anything previously imaginable.
The Parallel to Today
Similar to CAD, AI cannot create the experience and deep industry knowledge necessary to transform manufacturing. The technology is necessary but not sufficient.
This is precisely what the research on AI implementation success confirms: the most successful AI implementations involve partners who understand specific manufacturing processes, quality requirements, and operational constraints. Companies that successfully integrate domain knowledge into AI achieve dramatically better results than those deploying generic AI applications — the difference between actionable insights and expensive disappointments.
The companies currently failing at AI — the 74% seeing no tangible value, the 42% abandoning their initiatives — are often making the same mistake: treating AI as a product to be purchased rather than a capability to be cultivated. They're forgetting that CAD didn't win because it was CAD. CAD won in Germany because Germany was ready.
The Path Forward: A Lesson for Industry at the Dawn of a new Era
Today, Germany — along with industrial leaders worldwide — faces a remarkably similar moment. The AI technology is American. The compute infrastructure is American. The foundational models are American.
But the implementation? The domain expertise? The decades of industry knowledge encoded in engineers, processes, and institutional memory? The ecosystems of specialized excellence? The cultures of precision and craftsmanship built over generations?
That belongs to industry leaders around the world — whether German automotive engineers, Japanese manufacturers, Korean semiconductor specialists, or American aerospace companies. Each holds irreplaceable expertise in their domains.
Manufacturers implementing collaborative human-AI solutions achieve 3.7x ROI on their investments—but only when they structure AI to augment existing expertise rather than replace it. The future belongs not to those who build the best AI, but to those who best integrate AI into domains where they already excel.
Germany didn't invent CAD. Germany won with CAD.
The lesson is universal: in eras of transformative technology, the winners are not necessarily the inventors. They are the integrators — those with existing excellence, cultural readiness, and the strategic wisdom to recognize that new tools amplify existing strengths rather than replace them.
This is the moment for every industry leader sitting on decades of hard-won expertise. Not to chase Silicon Valley at its own game, but to do what industry has always done best: take powerful general-purpose technology and apply it with precision to domains where your expertise runs deepest.
For Germany, this means finding a way not just to survive, but to lead a new era — combining automotive and manufacturing mastery with AI that amplifies human expertise. For others, it means identifying where your decades of domain knowledge create an insurmountable advantage, then building AI systems that make that knowledge scale. The carefully cultivated networks of world-renown experts can be the deciding advantage.
The path forward requires bold strategic thinking and a trajectory beyond the role of paying customer. Not copying what tech players are doing or handing over control, but carving a different path: sovereign industry transformation that combines irreplaceable human knowledge with cutting-edge AI capabilities.
The world has changed. A new kind of AI is here. Every industry is changing, every information-based value-chain brings the potential to disrupt and win. In many industries we will see new empires being built. This transformation is just getting started.