Analysis
Siemens plans to invest more than €1 billion in industrial AI over the next three years, betting that Europe's century-old manufacturing base gives it a data advantage American and Chinese AI labs can't easily replicate, Fortune reports in an interview with Chief Technology Officer Peter Koerte. More than half of the companies on the Fortune 500 Europe list are over 100 years old, and Koerte's argument is that the decades of proprietary engineering, production and failure data sitting inside European pharma, automotive and chemical companies is exactly the training material general-purpose language models don't have.
Industrial AI Is A Different Bet Than Chatbots
Siemens' Industrial Foundation Model, still in development, is trained on manufacturing and engineering data rather than text, and Koerte says it could shorten engineering cycles by up to 40% in automotive and aerospace design work. The distinction matters because the failure mode is different and more expensive: a consumer chatbot hallucinating a fact is embarrassing, but Koerte told Forbes in January that an industrial model inventing a calculation in a production context is a safety and liability problem, not a PR one -- which is why Siemens is building narrower, domain-specific models instead of pointing a general LLM at factory data.
Siemens competes here against GE Vernova and Honeywell in industrial software, and against the hyperscalers -- Microsoft, Google and Amazon -- all separately pitching "industrial AI" platforms layered on top of their cloud infrastructure rather than built from decades of proprietary shop-floor data.
The risk in Koerte's framing is that it is also a convenient story for a European industrial giant to tell investors anxious about US and Chinese AI dominance -- a narrative critics could reasonably call more strategic positioning than settled fact, since Siemens has not yet shown the Industrial Foundation Model working at the scale or reliability the 40% engineering-cycle claim implies. Koerte himself favors collaboration over "technological isolationism," arguing cross-border cooperation, not a defensive European AI bloc, is what actually captures the efficiency gains -- a more measured claim than the headline-friendly version of European AI sovereignty now circulating in Brussels policy circles.