STM32 products are integrated into NPU, finally catching up with the AI competitive landscape (with an introduction to the entire STM32 series)
Under the wave of end-side AI, MCU integrated NPU has become standard.STMicroelectronics (ST) finally launchesSTM32N6 series, equipped with self-developed NPU, officially entered the AI MCU track, making up for the previous shortcomings in intelligent computing power.
1. STM32N6 makes a grand debut: AI computing power jumps 600 times
- NPU computing power: Equipped with ST self-developedNeural-ART NPU, up to 1GHz0.6 TOPS, the machine learning capability is that of the previous generation flagship STM32600 times
- Main control core：Arm Cortex-M55, the main frequency is up to0.8GHz
- Storage configuration: built-in4.2MB RAM, the highest specification in the history of ST MCU
- Function extension: Integration onlyISP, can directly process image signals
- Manufacturing process：16nmAdvanced manufacturing process supports high integration and low power consumption
With its powerful neural network reasoning capabilities, STM32N6 can cover end-side AI scenarios such as computer vision, audio processing, smart industry, smart home, automobiles, and consumer electronics.
2. The AI MCU market has become extremely hot, and it is difficult for ST to catch up from behind.
Before the release of STM32N6, many manufacturers had taken the lead in deploying AI MCUs:
- TI：TMS320F28P55x, known as the industry’s first real-time MCU with NPU
- NXP: RT700, i.MX 94 series are launched one after another
- Infineon：PSoC Edge series focuses on edge AI
- local manufacturers: Follow up the AI MCU solution one after another
ST’s core advantage still lies inComplete ecosystem, huge developer group, and product line covering all scenarios。
3. Sort out the entire STM32 series (understand the positioning in one article)
1. STM32F series - classic high-performance mainstream
- F0: low cost, low power consumption
- F1: The general main force in low-end and mid-range
- F4/F7/F8: high performance, with FPU and DSP, for industrial and high-end consumer electronics
2. STM32G series - cost-effective and general-purpose
- G0: Cortex-M0+, entry level
- G4: Cortex-M4, for motor control, industry, and home appliances
3. STM32L series - low-power battery-powered scenario
- L0/L1: ultra-low power consumption
- L4: High performance + low power consumption balance, suitable for wearable and IoT sensor nodes
4. STM32H series - flagship high performance
- H7/H5: Cortex-M7 core for computing-intensive embedded applications
5. STM32W Series - Wireless IoT MCU
- WB: Dual core, supports Bluetooth, Zigbee, Thread
- WBA: Cortex-M33, supports BLE5.4, Matter, BLE Audio
6. STM32MP series——MPU microprocessor
- MP2: 64-bit, Cortex-A35 + M33, integrated1.35TOPS NPU, oriented to complex scenarios such as edge AI, gateways, and industrial displays.
Other supplements
- STM32U0: ultra-low power entry model
- STM32H7R/S: high performance and cost-effective model
4. Summary
The launch of STM32N6 marks a key step for ST to officially complete **General MCU → AI MCU**.
With a complete series layout, mature ecosystem and huge customer base, ST is expected to quickly regain lost ground in the device-side AI competition and compete head-on with TI, NXP and Infineon.
$1.3/ 个
BOM
+86 4000004090
Online WhatsApp