Ambiq launched two new edge AI runtime options, HeliosRT and HeliosAOT, optimized for his or her Apollo SoCs to reinforce AI efficiency and vitality effectivity in edge computing.
HeliosRT is a power-optimized model of LiteRT (TensorFlow Lite for Microcontrollers) providing as much as 3x enhancements in inference velocity and energy effectivity. HeliosAOT is an ahead-of-time compiler that converts TensorFlow Lite fashions into embedded C code, lowering reminiscence utilization by 15–50% and enhancing deployment flexibility.
Each options tackle challenges in deploying AI on ultra-low-power gadgets like wearables, IoT sensors, and industrial displays. Constructed on Ambiq’s patented SPOT expertise, these instruments ship important energy consumption enhancements for edge AI purposes.
“The intersection of developer expertise and energy effectivity is our north star,” says Carlos Morales, VP of AI at Ambiq. “HeliosRT and HeliosAOT are designed to combine seamlessly with present AI growth pipelines whereas delivering the efficiency and effectivity positive aspects that edge purposes demand. We consider this can be a main step ahead in making refined AI actually ubiquitous.”
HeliosRT is offered in beta, with a basic launch anticipated in Q3 2025, whereas HeliosAOT is in technical preview for choose companions, with wider availability deliberate for This fall 2025.
The instruments combine seamlessly with present AI growth workflows and are supported by documentation, examples, and engineering help.
Associated
AI/ML | Ambiq | edge AI | Edge Inference | IoT
