Sensors Converge, the annual conference and expo in Santa Clara, remains important because it brings together engineers who deal not with abstractions about AI and other issues but with real signals, power budgets, latency limits, form factors, and product constraints.

Participants in the 2026 edition gathered to examine how intelligence integration is advancing further into the sensing stack.  The sensor industry is no longer just about measuring the physical world — it’s about understanding it. The conversation has moved to “how close to the sensor can we put intelligence?”

TDK InvenSense’s CTO, Omar Abed, declared 2026 a turning point for edge AI; Intel’s keynote highlighted software-defined sensor fusion for autonomous systems; STMicroelectronics promoted MEMS, imaging, wireless connectivity, and edge AI demos. Microchip highlighted local AI inference on embedded processors for applications such as access control, smart buildings, and embedded vision. Bosch Sensortec framed its presence around AI-powered sensing and real-world impact. 

The direction is crystal clear. Sensor nodes are no longer dumb transducers. They’re becoming perception engines.

Pete Bernard, Edge AI Foundation CEO, jokingly titled his address “Edge or Extinction.” Aside from the humor, the underlying point resonated: if your sensor only sends raw data upstream, you’re already behind.

Francesco Doddo, AIoT Sensors Applications Director for STMicroelectronics in the Americas, touched on a concept POLYN constantly promotes: the neuromorphic analog front-end as the next step in sensor intelligence.

The idea is simple and profound. Rather than handling a raw signal through traditional digital systems, we use brain-inspired processing directly in the analog domain, at the very moment the sensor produces the signal.

POLYN’s NASP technology operates exactly at this boundary, the analog frontier where physical signals can transform into meaningful events.

Major industry players are now building product roadmaps around this concept.

Highlights from the Sensors Converge Awards underscored the intelligence-everywhere trend. Not just smarter sensor nodes, but a different system architecture: one where intelligence is distributed, layered, and placed as close as possible to the physical source of information.  

On the expo floor, Toborlife AI‘s humanoid and quadruped robots drew the biggest crowds near the entrance, while Hexagon and Murata demonstrated safety-certified lane-level positioning for autonomous vehicles, a live showcase of MEMS, IMUs, and GNSS sensor fusion. UpbeatTech showcased a “Tiny AI Engine” built around small vibration sensors paired with a microprocessor, essentially on-device AI for wearables and predictive maintenance in a fingernail-sized package.

As sensor deployments scale, the industry is rethinking power sources. Energy harvesting is gaining serious momentum as the path to battery-free intelligent sensing.

The future of AI in physical systems will be decided by how efficiently machines can perceive the world at the source. The question is how fast the sensors will get smarter, and who builds the architectures that make it possible. POLYN is here for that conversation.