embedUR launches AI model platform for low-power devices
Fremont, California-based embedUR with major Chennai operations launched ModelNova—a platform packed with over 150 lightweight AI models for gadgets like wearables, smart cameras, sensors, and robots.
These models handle everything from object detection and face recognition to audio clean-up and spotting unusual activity.
They're built to run smoothly on low-power chips (think TensorFlow Lite Micro), so they're perfect for devices that need to be fast but energy-efficient.
ModelNova comes with Fusion Studio
ModelNova comes with Fusion Studio—a tool that lets developers tweak, test, and roll out these AI models way faster (up to 75% less time).
The platform provides performance metrics to help evaluate model size, accuracy, and estimated power consumption.
The platform makes it easier to build cool stuff in fields like AR/VR, healthcare tech, smart farming, or even industrial automation.
embedUR is investing ₹100 crore into a new edge intelligence center
To make ModelNova work on more hardware, embedUR is teaming up with companies like Ceva (for NeuPro NPUs), Infineon, STMicroelectronics.
On top of that, they will invest ₹100 crore into a new edge intelligence center in Chennai—part of a bigger ₹500 crore plan—to hire 100 engineers focused on next-gen telecom and semiconductor projects.