114991684

Overview

The Seeed Studio SKU 114991684 is the Sipeed MAIX-I module WiFi version, an AIoT edge computing module based on the Kendryte K210 RISC-V processor. It integrates a dual-core 64-bit CPU with KPU and APU hardware accelerators for neural network and audio processing. The module includes an ESP8285 Wi-Fi chip, DC-DC power regulation, and selectable flash memory options of 8MB, 16MB, or 128MB. Operating at 400MHz (overclockable to 800MHz), it delivers up to 0.5 TOPS performance within a 25mm x 25mm x 1mm form factor.

Feature Summary

• Kendryte K210 RISC-V 64-bit dual-core CPU with 8MB on-chip SRAM
• Integrated ESP8285 Wi-Fi chip for wireless connectivity
• Hardware accelerators: KPU (up to 0.5 TOPS @ 800MHz) and APU
• Flexible FPIOA mapping 255 functions to 48 GPIOs
• Supports DVP camera and MCU LCD interfaces

Applications

• Edge AI inference and machine vision
• Voice recognition and microphone array processing
• Smart IoT devices requiring low latency and privacy
• Robotics and autonomous systems

Procurement Notes

Verify the specific flash capacity configuration (8MB, 16MB, or 128MB) as the datasheet lists multiple options without specifying the exact quantity for this SKU. Confirm pin voltage selection capabilities (1.8V or 3.3V) against the target system requirements before integration.

Selection Notes

Select this module when a compact footprint is required for edge AI tasks. Ensure the application leverages the integrated Wi-Fi capability via the ESP8285. Consider the overclocking potential to 800MHz if higher throughput is needed, noting the associated power increase.

Part Context

This module is part of the Sipeed MAIX product line designed for edge computing. It serves as the core component for development boards such as the M1 Dock, M1w Dock, and Maix Bit. The WiFi variant specifically adds wireless communication capabilities compared to standard modules.

Replacement Considerations

Publicly confirmed substitute part numbers are not available in the provided documentation. Alternative solutions would require evaluating other RISC-V AI modules with similar KPU/APU architectures and form factors.