The new chips feature a specialized sensing hub capable of running models with up to 200 million parameters. This architecture supports local tasks like speech differentiation and personal scribing, allowing devices to build a user-specific memory bank for task automation. The Extreme version significantly expands these boundaries, supporting a 30-billion-parameter mixture-of-experts model that activates only necessary parameters per task, surpassing the 20-billion-parameter capacity Apple showcased at its recent developer conference.
Beyond processing power, the chips integrate advanced camera controls and an APV codec to support 8K video recording at 60 frames per second. AI-driven features extend to audio as well, with new voice-isolation technology designed to strip away background noise during calls. Motorola has already confirmed the integration of the Extreme chip into its upcoming Signature 27 smartphone. This hardware push aligns with a broader industry shift toward mobile devices as the primary interface for AI, a sentiment recently echoed by leadership at both Nothing and Apple.

Comments (0)
No comments yet. Be the first!