AIStorm and Honda Partner to Reduce AI Hardware Power Drain

AIStorm Inc is implementing hardware strategies to address the energy inefficiency of data movement in AI processing. CEO David Schie argues that moving data off-chip can consume 100 to 1,000 times more energy than the actual arithmetic operations, often making devices undeployable due to battery drain.

To mitigate this, AIStorm couples AI processing directly to raw analog output to maintain power consumption in the single-digit microwatts. This approach aligns with industry shifts toward moving computation closer to sensors, such as Sony's IMX500 image sensor, which runs inference on the sensor die, and Prophesee's event-based vision sensors.

Expanding these efforts into the automotive sector, Honda Motor Company signed a joint development agreement in 2026 to co-design an AI chip. The new chip integrates parameter storage directly into the compute layer to minimize the energy cost of data shuttling.