industry news
Subscribe Now

Ambient Scientific Announces GPX-10 Revolutionizing AI in Endpoint and Edge Devices

Ambient Scientific introduces GPX-10, industry’s lowest power Artificial Intelligence (AI) general purpose  programmable processor. GPX-10 with “Always-on” Ambilogic technology is designed to enable AI in battery  operated and other fast growing low power devices. GPX-10 is the first of a rich roadmap of programmable AI  processors enabling inference and training capabilities.

GPX-10 is ideally suited to provide AI capabilities for vision applications such as Presence detection, Gesture  recognition, Image recognition, computer vision and for voice applications such as Custom Wake words, voice  recognition and authentication. With these capabilities, GPX-10 is well suited for a variety of markets which include  client-side devices, healthcare, automotive, gaming, white goods, IOT and industrial products.  

Ambient Scientific’s founder and CEO, GP Singh, said, “GPX-10 can bring AI inference capabilities to the client- side  applications at low-power and cost-effectively without the need for backend cloud support enabling solutions that  will allow fast local decision making for a broad range of power-sensitive connected and standalone unconnected  devices and appliances.”  

GPX-10, with its ten neural network cores and ARM Cortex-M4 host controller provides 400KB of SRAM onboard and  an external flash memory interface. The SoC provides 8 analog inputs and a separate input for vision sensors.  Ambient’s “always-on” logic consumes less than 100 microwatts of power, and in full operation, at 100MHz less than  300 microwatts of power for most voice applications. GPX-10 can be used for image recognition and in combination  with voice or other analog inputs, it enables concurrent multisensor-based AI applications.  

Ambient is announcing availability of the Development kits and Software libraries for voice and vision applications.  Ambient’s compiler allows high-level Python based code development that will allow fast time to market for  customers. GPX-10 processor, now sampling, is available either in die form or in custom packages.

About Ambient Scientific  

Ambient Scientific is a leading developer of industry’s lowest power programmable AI processors designed to  address the explosive demand for Inference and Training in edge and Battery-operated AI devices giving both  connected and unconnected devices and appliances their own personalities. Ambient’s AI processor and  microcontroller products and software libraries are designed to bring a plethora of exciting new innovative products  to the market quickly.  

By enabling AI at the edge, Ambient’s technology plays a key role in sustainable computing by alleviating data traffic  across networks and reducing the demand on the cloud datacenters which today power most of the AI inference  and training resources.  

Learn more at https://ambientscientific.ai/ Source: Ambient Scientific

Leave a Reply

featured blogs
May 6, 2026
Hollywood has struck gold with The Lord of the Rings and Dune'”so which sci-fi and fantasy books should filmmakers tackle next?...

featured paper

Quickly and accurately identify inter-domain leakage issues in IC designs

Sponsored by Siemens Digital Industries Software

Power domain leakage is a major IC reliability issue, often missed by traditional tools. This white paper describes challenges of identifying leakage, types of false results, and presents Siemens EDA’s Insight Analyzer. The tool proactively finds true leakage paths, filters out false positives, and helps circuit designers quickly fix risks—enabling more robust, reliable chip designs. With detailed, context-aware analysis, designers save time and improve silicon quality.

Click to read more

featured chalk talk

GaN for Humanoid Robots
Sponsored by Mouser Electronics and Infineon
In this episode of Chalk Talk, Eric Persson and Amelia Dalton explore why power is the key driver for efficient and reliable robot movements and how GaN technologies can help motor control solutions be more compact, integrated and efficient. They also investigate the role of field-oriented control in humanoid robotic applications and why the choice of a GaN power transistor can make all the difference in your next humanoid robot project!
Apr 20, 2026
17,165 views