Texas Instruments Launches Edge-AI MCUs with TinyEngine NPU to Bring Intelligent Processing to Everyday Embedded Systems

Texas Instruments  has introduced two new microcontroller families with built-in edge AI capabilities, reinforcing its strategy to bring AI across its entire embedded processing portfolio. The MSPM0G5187 and AM13Ex MCUs integrate TI’s TinyEngine neural processing unit (NPU)—a dedicated hardware accelerator tailored for microcontrollers that optimizes deep-learning inference, reduces latency, and improves energy efficiency for edge computing.

 

TI’s embedded processors are supported by a mature development ecosystem, including the Code Composer Studio (CCStudio) integrated development environment. With generative AI features and industry-standard models enriched by TI data, engineers can accelerate coding, system configuration, and debugging using natural language. These innovations enable edge AI across a wide spectrum of applications—from wearable health monitors and smart circuit breakers to physical AI in humanoid robots. TI will showcase these advancements at **embedded world 2026 in Nuremberg, Germany, from March 10–12, 2026.

 

According to Amichai Ron, Senior Vice President of Embedded Processing and DLP® at TI, the company’s early work in digital signal processing laid the groundwork for today’s edge AI evolution. By embedding the TinyEngine NPU into both general-purpose and high-performance real-time MCUs—and aligning devices, tools, and software—TI is making edge AI practical and accessible for nearly any application.

 

Bob O’Donnell, President and Chief Analyst at TECHnalysis Research, noted that while much attention is focused on large SoCs with AI acceleration, some of the most impactful AI use cases can be realized in small, efficient microcontrollers. When paired with AI-assisted software tools, these chips can bring intelligent capabilities to a far broader range of engineers and device designers.

 

Bringing advanced intelligence to everyday devices

From fitness wearables to household appliances and electrical systems, consumers increasingly expect smarter devices. Historically, AI functions were limited to premium applications due to cost, power, and complexity constraints. The new MSPM0G5187 Arm® Cortex®-M0+ MCU changes this dynamic, enabling designers to embed edge AI into compact, affordable, and energy-efficient systems.

 

The TinyEngine NPU operates in parallel with the main CPU, offloading neural-network computations while application code runs uninterrupted. Compared to MCUs without hardware acceleration, this architecture can:

• Reduce flash memory requirements
• Lower inference latency by up to 90×
• Cut energy consumption per inference by more than 120×

These efficiencies make AI workloads feasible even for battery-powered and resource-constrained devices. With pricing below $1 per unit in volume, the MSPM0G5187 provides a cost-effective alternative to more complex processor architectures.

 

Real-time motor control combined with AI acceleration

Modern appliances, robotics, and industrial equipment increasingly require intelligent motor control features such as adaptive control and predictive maintenance. Previously, this required multi-chip solutions. Building on more than two decades of leadership in motor control MCUs, TI’s AM13Ex MCU is the first to combine a high-performance Arm Cortex-M33 core, the TinyEngine NPU, and advanced real-time control on a single chip.

 

This integration reduces bill-of-materials costs by up to 30% while enabling:

• Precise real-time control of up to four motors alongside adaptive AI algorithms for load sensing and energy optimization
• An integrated trigonometric math accelerator delivering calculations 10× faster than traditional CORDIC implementations for improved control responsiveness

Simplifying AI model development and deployment

To support these MCUs, TI offers CCStudio Edge AI Studio, a free environment for model selection, training, optimization, and deployment across TI’s embedded processors. The toolchain currently includes more than 60 models and application examples, allowing developers to quickly implement edge AI in diverse devices, with additional models planned.

 

At embedded world 2026, TI will demonstrate how its technologies help engineers accelerate development with AI, enhance performance through edge intelligence, and deploy AI at the edge in factories, buildings, vehicles, and beyond.