The SDK embeds the LiteRT runtime in the modules’ firmware, which runs AI inference on their application cores 

• Combines AI inference and cellular connectivity in a single module 
• Supports the standard .tflite model format without a vendor-specific rebuild  
• Includes sample applications and generic reference models 

Boca Raton, Fla. — Telit Cinterion, an end-to-end Internet of Things (IoT) solutions enabler, today announced both its edge AI SDK and planned AI-enabled cellular module variants. The SDK runs machine learning models on a module’s own hardware without the need for an external AI accelerator or companion processor. 

Expected in Q4 2026, the edge AI SDK incorporates LiteRT, formerly TensorFlow Lite, into the Linux-based firmware of upcoming Telit Cinterion AI-enabled cellular modules. Those module variants will span 4G, 5G RedCap and high-performance 5G. 

LiteRT runs the standard .tflite format. Developers train and optimize models in established tools, then load the resulting file onto a Telit Cinterion module without rebuilding for a proprietary environment. A model already running on a PC or a Raspberry Pi transfers directly. 

The compact runtime fits on small, low-cost modules unable to host larger AI software stacks. That footprint extends on-module inference to hardware previously limited to connectivity. 

Sample applications included in the SDK move a model through the full inference pipeline, acquiring sensor data, preprocessing it, running the model and returning the prediction to the industrial application. System integrators retain control over the final model, application logic and deployment architecture. 

In proof-of-concept testing on image classification and object detection, inference consumed no more than 17% of CPU resources. That margin holds the module below the point where thermal throttling would degrade 4G and 5G performance. 

Potential customer-developed applications include: 

• Predictive maintenance: Analyze vibration or audio data from motors, pumps and bearings to identify anomalies before equipment fails.  

• Acoustic monitoring: Recognize alarms, breaking glass and other important sound patterns at remote industrial or infrastructure sites.  

• Smart metering: Read existing analog meters through connected cameras and image classification, with no replacement of installed equipment.  

“Industrial IoT teams should not have to redesign their entire device architecture to add practical AI capabilities,” said Marco Argenton, senior vice president of product management at Telit Cinterion. “By bringing a lightweight AI runtime into the cellular module, we help customers reduce system complexity and accelerate the path from proof of concept to a connected industrial solution that can operate reliably in the field.” 

“The real engineering challenge isn’t simply running AI at the edge. It’s delivering secure, efficient and reliable intelligence within the constraints of a connected embedded device,” said Vishal Batra, vice president of software engineering at Telit Cinterion. “Our edge AI SDK enables developers to deploy standard AI models directly on Telit Cinterion modules, accelerating development without compromising the trusted connectivity industrial IoT applications require.” 

For additional information about the Telit Cinterion edge AI SDK and planned AI-enabled cellular modules, visit edge AI SDK or contact Telit Cinterion.