STMicroelectronics is ushering in the era of edge intelligence in vehicles with the introduction of its first automotive microcontroller (MCU) with built-in AI acceleration.

ST announced on the 24th that it is launching Stellar P3E, the next-generation automotive electronics architecture for software-defined vehicles (SDVs).

The Stellar P3E is an MCU designed to run real-time edge AI applications directly inside the vehicle, simplifying X-in-1 Electronic Control Unit (ECU) integration to reduce system cost, weight, and complexity.

Provides safe and responsive real-time control performance from hybrid and electric vehicle (EV) systems to vehicle body zone architecture.

The key to this product is the integration of ST’s Neural-ART Accelerator™, the automotive industry’s first embedded neural network accelerator (NPU).

This enables fast inference speeds in microseconds and provides AI efficiency that is up to 30 times improved compared to existing MCUs.

Implementing low-power AI capable of always-on operation enables various vehicle intelligence functions such as predictive maintenance, virtual sensors, and smart sensing. Support.

Additionally, the Stella P3E is equipped with ST's phase-change memory (PCM)-based non-volatile memory 'xMemory', which allows for flexible expansion of software storage space.

This helps add new features without hardware redesign in SDV environments where software updates and feature expansions are frequent.

“The Stellar P3E sets a new standard for automotive electrification by integrating high-performance real-time control and edge AI in a single device,” said Luca Lodeschini, Vice President of STMicroelectronics. “It will help automakers deliver safer and more efficient driving experiences.”

Greg Vasich of Counterpoint Research commented, “Moving AI computation to the edge of the vehicle enables sub-millisecond decision-making,” adding, “High-precision sensing and control can be implemented at lower cost and heat generation compared to full-scale SoCs.”

The Stellar P3E supports the entire process from data generation to on-device deployment through the ST Edge AI Suite and Stellar Studio, and is scheduled for mass production in Q4 2026.