Roland Bush simplifies the complex, 170-year-old Siemens by framing it not by its products, but by its core function: providing the underlying technology that enables other companies to operate and innovate in industries from automotive to healthcare.
The CEO of Siemens advocates for decisions to be made at the lowest possible level. However, he stresses this empowerment is a two-way street that must operate within clear strategic boundaries and come with direct accountability for the outcomes, preventing chaos.
To transform the 320,000-person company, Siemens' leadership avoided a top-down restructuring mandate. Instead, they defined a clear "North Star" vision and then empowered employees to co-create the "tracks" (initiatives) to reach it, fostering broad buy-in and ownership.
Siemens mitigates geopolitical risks and tariffs not just by being global, but by being hyper-local. Its CEO reveals that 85-87% of its production in major markets like the US and China is for that market, minimizing cross-border dependencies and the direct impact of trade wars.
Siemens navigates its immense scale through a three-dimensional matrix of businesses, regions, and industry verticals. Critically, the primary axis of power and P&L responsibility lies with the global business units, not geography, though this model adapts for certain divisions.
Roland Bush asserts that foundational LLMs alone are insufficient and dangerous for industrial applications due to their unreliability. He argues that achieving the required 95%+ accuracy depends on augmenting these models with highly specific, proprietary data from machines, operations, and past fixes.
Siemens discovered that standard virtual training for robots was insufficient for real-world application. The robot's accuracy only jumped to a usable level after they switched to a photorealistic digital twin using advanced ray-tracing, which more accurately modeled light and texture for the AI.
To break down rigid business units, Siemens' CEO is creating horizontal "fabrics" for data, technology, and sales. These thin but powerful layers act like a shared operating system to enforce standards and scale capabilities across the entire organization without a full functional re-org.
Responding to fears of job loss from automation, Siemens' CEO frames it as a necessary shift. In aging societies with labor shortages, automating manufacturing allows for economic growth while redeploying human workers to essential, non-automatable sectors like healthcare and social services.
To overcome the data scarcity problem for industrial AI, Siemens formed an alliance with competing German machine builders. These companies agreed to pool their operational data, trusting Siemens to build powerful, shared AI models that are more effective than any single company could create alone.
Recognizing that employees in less glamorous but profitable divisions (like mechanical switches) can feel ignored, Siemens' CEO actively works to validate their contribution. He connects their work directly to customer value and the company's financial health, ensuring they don't feel lost in the AI hype.
Contrary to political rhetoric, Siemens' CEO provides a ground-level view that a widespread return of manufacturing to the US has not yet materialized. He cites labor shortages and policy uncertainty as key drags, despite real investments in specific sectors like pharma and semiconductors.
Rather than trying to predict specific geopolitical crises, Siemens builds resilience by creating separate technology stacks for different regions. For instance, its industrial AI for China is trained on Chinese LLMs, while its US counterpart uses American models, creating independent and compliant systems.