TLDR: Renesas Electronics has launched the RA8P1 microcontroller (MCU) group, setting a new performance benchmark for edge and endpoint AI applications. These MCUs integrate 1GHz Arm Cortex-M85 and 250MHz Cortex-M33 CPU cores with an Arm Ethos-U55 Neural Processing Unit (NPU), delivering over 7300 CoreMarks and 256 GOPS AI performance. Built on TSMC’s 22ULL process, the RA8P1 features embedded MRAM, extensive memory options, and advanced security, making it ideal for vision, voice, and real-time analytics in AIoT devices.
TOKYO, Japan – Renesas Electronics Corporation (TSE:6723), a leading provider of advanced semiconductor solutions, has officially introduced its groundbreaking RA8P1 microcontroller (MCU) group, specifically engineered to accelerate Artificial Intelligence (AI) and Machine Learning (ML) applications directly at the edge. Unveiled on July 1, 2025, this new series establishes a significant leap in MCU capabilities, combining high-performance processing with dedicated AI acceleration.
At the core of the RA8P1 MCUs are a powerful 1GHz Arm Cortex-M85 CPU and a 250MHz Cortex-M33 CPU, complemented by the integrated Arm Ethos-U55 Neural Processing Unit (NPU). This formidable combination achieves a CPU performance exceeding 7300 CoreMarks and an impressive AI performance of 256 GOPS (Giga Operations Per Second) at 500 MHz. The Ethos-U55 NPU is particularly noteworthy, capable of delivering up to 35 times more inferences per second compared to the Cortex-M85 processor operating independently, by offloading compute-intensive operations for Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). It supports widely used networks such as DS-CNN, ResNet, MobileNet, and TinyYolo.
Built on TSMC’s advanced 22ULL (Ultra-Low Leakage) manufacturing process, the RA8P1 group offers both high performance and low power consumption. The MCUs feature robust memory options, including 2MB of on-chip SRAM for intermediate activations and graphics framebuffers, and 1MB of embedded Magnetoresistive RAM (MRAM). The MRAM provides superior write speeds, higher endurance, and better retention than traditional flash memory, making it ideal for storing application code and AI model weights. For more demanding AI applications, SIP (System-in-Package) options with 4MB or 8MB of external flash are also available, alongside high-speed external memory interfaces.
The RA8P1 is optimized for a wide array of edge and endpoint AI applications, including machine vision, smart surveillance, video doorbells, face recognition panels, smart speakers, predictive maintenance, anomaly detection, and industrial automation. Its integrated graphics capabilities, MIPI-CSI2 and CEU camera interfaces, and a broad range of peripherals like I2S, PDM, and Gigabit Ethernet with TSN, further enhance its versatility for vision and voice AI, as well as real-time analytics.
Security is a paramount feature of the RA8P1 MCUs. They incorporate Renesas’s proprietary Security IP (RSIP-E50D) hardware security module, which includes cryptographic accelerators for algorithms such as CHACHA20, Ed25519, NIST ECC curves, enhanced RSA, SHA2, and SHA3. The devices also provide a strong hardware Root-of-Trust and Secure Boot with a First Stage Bootloader (FSBL) in immutable storage. XSPI interfaces with decryption-on-the-fly (DOTF) allow encrypted code images to be stored externally and securely executed.
Daryl Khoo, Vice President of Embedded Processing Marketing Division at Renesas, emphasized the significance of this launch, stating, “There is explosive growth in demand for high-performance edge AIoT applications. We are thrilled to introduce what we believe are the best MCUs to address this trend. Customers are eager to employ these new MCUs in multiple AI applications.” Khoo also highlighted that the chip represents ‘the next generation of Renesas RA microcontroller family, designed for the age of AIoT,’ stressing the importance of efficient, real-time, and cost-effective local AI processing at the extreme edge.
Also Read:
- Ambiq Apollo510 SoC Integrates with Edge Impulse Platform, Revolutionizing Efficient Edge AI Development
- Ambiq Micro Seeks U.S. Public Listing Amid Surging AI Chip Demand
Renesas has also invested in easing the development experience for customers, recognizing the challenges in AI development. The RA8P1 MCUs are available now, poised to enable a new generation of intelligent, secure, and highly efficient embedded systems.
