Upbeat Technology, in collaboration with SiFive, has introduced UP201 and UP301 always-on AI MCUs for ultra-low-power AI and IoT applications such as wearables, drones, and sensor-based systems. The UP201 is designed for compact, battery-driven devices such as smartwatches, hearing aids, and IoT sensor nodes, whereas the UP301 targets AI and vision-based applications for more complex edge systems such as smart glasses, robotics, and industrial AI equipment.
Both chips feature a dual-core RISC-V architecture. The SiFive E21 lightweight core is Always-On (AON), managing continuous low-power sensing, whereas the SiFive E34 performance core activates for higher workloads. The MCUs also integrate a 2.5D GPU and various I/O options.
UP201 and UP301 MCU specifications:
- CPU cores (SiFive Essential RISC-V IP cores)
- E21 core in Always-On (AON) domain for continuous low-power sensing
- E34 performance core in the Non-AON domain for higher workloads
- Up to 400 MHz operating frequency
- Up to 717 DMIPS performance
- Near-threshold operation for ultra-low active and sleep power consumption
- AI accelerator
- NPU1 for complex AI workloads (Speaker ID, ECG, PPG, etc.)
- NPU2 (Always-On) for keyword or gesture detection
- Dual NPUs (NPU1 + NPU2)
- Dedicated DSP acceleration for signal processing and sensor fusion
- Memory
- On-chip SRAM with Upbeat’s Error Detection and Correction architecture (capacity not specified)
- External memory expansion supported through SPI interfaces
- Storage
- Quad SPI Flash interface for UP201
- Built-in SDIO 3.0 and eMMC 4.5.1 support for UP301
- Graphics – Integrated GFX 2.5D GPU
- Display
- UP201 – SPI (DBI-C) display interface
- UP301 – 2-lane MIPI-DSI-2 and QSPI display support
- Camera – 8-bit parallel camera interface for UP301
- Expansion
- Up to 29x GPIO, 3x SPI, 3x UART, 10x I2C for UP201
- Up to 100x GPIO, 6x SPI, 5x UART, 12x I2C for UP301
- Power Consumption – 16.8 uW/MHz/DMIPS (Dhrystone)
- Dimensions
- UP201 – 3mm × 3mm WLCSP, 64 balls
- UP301 – 5mm × 5mm FCCSP, 152 balls
According to the company, this dual heterogeneous core architecture enables near-instant wake-up and reduces energy use by up to 65 percent compared to conventional microcontrollers.
The company also designed the Trina-pi, a compact 61 x 31 mm development board based on the UP201 MCU. The development board includes the UP201 MCU, FT2232HP USB bridge, NOR Flash, power management ICs, GPIO headers, and reset and programming pushbuttons. The USB Type-C interface is used for power and programming, and the company is supposed to provide SDKs and examples, videos for neural inference, control tasks, display rendering, and much more. But after searching for quite a bit, I was not able to find any sources for those claims.
Engineering samples are available for the UP201/UP301 series always-on AI MCUs, and you have to get in touch with the company through a Google Doc form to get some. More information can be found on the product page and the press release.
Debashis Das is a technical content writer and embedded engineer with over five years of experience in the industry. With expertise in Embedded C, PCB Design, and SEO optimization, he effectively blends difficult technical topics with clear communication
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
