Theme:

Edge AI & Intelligent Systems

In 15 minutes from training to inference on the edge

Run the same AI workload on two architectures: a conventional MCU and a neuromorphic processor. Measure energy per inference, latency, and idle consumption under identical conditions.

Power traces must be recorded and compared transparently.

How to win the challenge?

• Lowest energy per inference

• Best performance-per-watt ratio

• Clear, validated power measurements

Bonus: Demonstrate advantage under event-driven workloads.

Participants: