Physical AI insights: FLEXOO CEO on Startup Insider podcast
2026-05-06
Physical AI and the limits of purely digital approaches
In a recent episode of the Startup Insider Podcast, FLEXOO CEO Michael Kröger shared his perspective on why Physical AI is becoming increasingly relevant as AI systems move from digital domains into real‑world physical environments.
In the conversation, Michael Kröger explained that the key limitation when applying AI to physical systems is often not model capability, but access to reliable physical data. Without direct sensing of parameters such as pressure, temperature, or other internal signals, AI systems remain limited in their ability to understand and act on physical processes.
Why physical data matters for real-world systems
The podcast discusses how many physical systems today are under‑instrumented, with critical changes occurring inside systems long before they become visible. As a result, AI models are frequently forced to infer system behavior from incomplete or delayed information.
This lack of direct access to physical signals creates a fundamental bottleneck for AI applications in safety‑critical and industrial environments.
Batteries as an early application domain
The conversation also highlights why batteries and energy systems are among the first application domains where Physical AI becomes tangible. According to Michael Kröger, safety‑critical behavior such as thermal runaway as well as reliable capacity determination at cell level require data that is typically not captured today.
This makes sensing a foundational element for applying AI meaningfully in these systems.
Edge intelligence and data volume
Another focal point of the discussion is data volume. Physical sensing can generate large, highly redundant data streams, which makes centralized processing inefficient. The podcast emphasizes the need for intelligence close to the sensor — filtering and identifying relevant events locally in order to enable timely reactions in real systems.
FLEXOO’s perspective on Physical AI
Overall, the episode reflects FLEXOO’s broader positioning in the Physical‑AI ecosystem: enabling access to real‑world physical data and closing the gap between sensing and action in applications where purely digital approaches reach their limits.
🎧Further reading and listening
Read the full recap and key takeaways from the conversation in our blog: When AI needs to feel the physical world
Listen to the Startup Insider Podcast featuring Michael Kröger (German language): https://startup-insider.simplecast.com/episodes/physical-ai-flexoo-sensortechnologie-series-a
Michael Kröger, CEO of FLEXOO, in conversation on the Startup Insider Podcast.