The Astra edge AI chip, developed by Google Research in partnership with Synaptics, was unveiled on Wednesday (October 14) at the Research@ event in Mountain View. The processor, based on the Coral NPU M2, cuts about 40% of power consumption compared to previous models, enabling always-on AI on wearables and IoT devices.
What is edge AI?
Edge AI processes data locally on the device without the cloud. This reduces latency, increases privacy, and allows quick responses for voice, video, and sensors. The big challenge is low power and heat in compact gadgets like glasses, watches, and smart earbuds.
Avinatan Hassidim, vice president of Google Research, explained that traditional edge chips were too slow and power-hungry for continuous use. The new Coral NPU M2 solves part of that.
Partnership and next steps
Synaptics has already integrated Google's open-source Coral NPU into the Astra SL2610 line and plans to demonstrate the chip in detail on October 20. CEO Rahul Patel stated that the technology will equip Google devices, although without specifying models or dates.
Ideal for processing voice, motion, and context in real time on wearables. Official sources and reports from Seoul Economic Daily and Chosun Biz.
By GeekikiBot
