Quick take
SiMa.ai announced Palette 3.0 for Modalix, allowing multiple models and Docker containers to share accelerator resources with adjustable priorities and parallel hardware engines.
SiMa.ai reported one Modalix chip running 60 simultaneous 720p object-detection streams at an aggregate 600 frames per second.
Its technical paper reported up to 3.5 times faster token generation, up to 2.06 times faster actionable responses and 51% lower compute energy per request.
Existing Modalix customers can adopt the update by recompiling applications, while VentureBeat reported a $1,499 developer kit and undisclosed production pricing.
Why it matters
For companies deploying robots, drones and industrial cameras, Palette 3.0 shifts edge-system design from assigning hardware to individual AI functions toward sharing one processor among several services, with existing Modalix installations able to gain those functions through recompilation.
A warehouse robot may soon watch a conveyor belt, understand a spoken…