March 18-19, 2027 · Munich, Germany · In person only
Submission deadline: Friday, 18 December 2026
Edge AI is moving from pilots into production, and increasingly into machines that sense and act in the physical world. EDGE AI Munich 2027 is where the engineers, researchers, and product teams doing this work compare notes. Munich is home to much of Europe’s automotive, industrial, and semiconductor industry, and the program reflects that.
We’re looking for talks grounded in real work: deployments, benchmarks on real hardware, tools you’ve built, and research with results. Lessons from what didn’t work are just as useful as what did.
The event is in person only. We don’t accept remote or pre-recorded talks.
Categories
This year we’re especially interested in the themes below. They’re a guide, not separate tracks: any edge AI topic is in scope, and abstracts that cover more than one theme are welcome.
Edge AI in Production
Real deployments and case studies; scaling from pilot to fleet; measured impact and lessons learned, including failures
Physical AI and Sensing
Robots, drones, vehicles, and machines that perceive and act in the real world; real-time perception and control; smart sensors and sensor fusion
Generative and Agentic AI
Small language and multimodal models on device; agents and tool use; hybrid edge–cloud inference
Hardware and Accelerators
NPUs and accelerators; neuromorphic and in-memory compute; RISC-V and chiplets; hardware–software co-design
Models, Tools and Frameworks
Compression and quantization; on-device and federated learning; compilers and runtimes; edge MLOps; benchmarks and interoperability
Energy and Sustainability
Measuring energy per inference; quantified savings of edge vs. cloud; energy-aware design and battery-free devices
Security, Safety and Regulation
Model and IP protection; privacy; adversarial robustness; functional safety; EU AI Act and Cyber Resilience Act in practice