Semiconductor and embedded processor maker Texas Instruments Inc. today announced the launch of new microcontrollers with TinyEngine neural processors.

The company said it’s introducing two new microcontroller unit, or MCU, families with edge artificial intelligence capabilities to support AI for embedded device platforms, such as smart watches, sensors and other connected devices.

The move adds to the company’s lineup of artificial intelligence-enabled hardware and software tools that allow engineers to deploy intelligence at the edge.

The MSPM0G5187 and AM13Ex MCUs use TI’s TinyEngine neural processor, which speeds up AI, reduces delay and cuts energy use.

“TI invented the digital signal processor almost 50 years ago, laying the groundwork for today’s edge AI processing,” said Amichai Ron, senior vice president, embedded processing and DLP products at TI. “By enabling AI across our software, tools, devices and ecosystem, we are making edge AI accessible and easy to use for every customer and every application.”

Artificial intelligence is winding its way across every part of the network experience, from the cloud to consumer devices. It also exists in industrial settings where AI does processing right at the edge, in sensors and machinery, before signals are sent upstream.

Every day, technology is becoming increasingly intelligent, as seen in fitness wearables, smartphones, thermostats, home appliances and cars. Other sectors are beginning to see intelligent technology become more prevalent as well, including industrial machinery, traffic lights, drones and farm equipment equipping smart sensors.

However, engineers still face the quintessential problem that AI is power-hungry. Many still believe AI should run in the cloud or on powerful hardware to handle large amounts of data and high-end applications. TI said its new Arm Cortex-driven MPUs can manage embedded designs that bring edge AI to a wider number of smaller, cost-effective applications.

Using local computation, the TinyEngine NPU provides neural networks in parallel to the primary central processor, which can handle the application code. Compared to similar MCUs without an accompanying accelerator, it can minimize memory footprint, lower latency by 90 times for AI inference and reduce energy utilization by more than 120 times.

Both new MCU families support TI’s CCStudio Edge AI Studio, a free development environment for developers that simplifies AI and machine learning model selection, training and deployment. It’s an edge AI toolchain for engineers with flexible AI model choices, including 60 models and application examples.

The company said production quantities of MSPM0G5187 MCU are available for purchase now, with the AM13E23019 MCU available in preproduction quantities. Variants of both with changes to package and memory sizes will be released by the end of 2026.

Image: Pixabay