{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigwbgntacgptt5vxupydr5k6louri5fczftugsojnfkenkjlygvrm",
    "uri": "at://did:plc:dmpmpvgbtnjd7pz4pxvvcwle/app.bsky.feed.post/3mgds3r4fthp2"
  },
  "path": "/2026/03/06/stm32u3b5-c5-ultra-low-power-mcu-features-640-kb-ram-2-mb-flash-and-hsp-accelerator-to-run-ai-without-batteries/",
  "publishedAt": "2026-03-05T17:01:08.000Z",
  "site": "https://www.cnx-software.com",
  "tags": [
    "Hardware",
    "STMicro STM32",
    "development board",
    "edge ai",
    "embedded world 2026",
    "energy harvesting",
    "low power",
    "mcu",
    "solar",
    "stm32",
    "stmicro",
    "STM32U3B5/C5 ultra-low-power MCU features 640 KB RAM, 2 MB Flash, and HSP accelerator to run AI without batteries",
    "CNX Software - Embedded Systems News"
  ],
  "textContent": "STMicroelectronics has added two members to the STM32U3 ultra-low-power Arm Cortex-M33 microcontroller family: the STM32U3B5 and STM32U3C5 MCUs get more sources with up to 640 KB SRAM and 2 MB flash, as well as an HSP (hardware signal processor) accelerator to run AI/ML workloads without batteries, just using energy harvesting. The new chips are still clocked at up to 96 MHz, benefit from a near-threshold design (down to 0.65 V), allowing a power consumption of just 117 Coremark/mW in active mode, and can operate up to 105°C ambient temperature. They come with one additional group of interfaces, bringing the total to four SPI and I2C, two I3C and CAN-FD, and five UARTs, as well as five more 16-bit timers, for a total of 10. The STM32U3C5 also includes a cryptocore to accelerate encryption and decryption operations, and supports the CCB (Coupling and Chaining Bridge) hardware security feature, both of which [...]\n\nThe post STM32U3B5/C5 ultra-low-power MCU features 640 KB RAM, 2 MB Flash, and HSP accelerator to run AI without batteries appeared first on CNX Software - Embedded Systems News.",
  "title": "STM32U3B5/C5 ultra-low-power MCU features 640 KB RAM, 2 MB Flash, and HSP accelerator to run AI without batteries"
}