# BeagleBone AI-64

> Powerful open-source AI/ML single board computer featuring TI TDA4VM SoC with dual Arm Cortex-A72 at 2.0 GHz, 8 TOPS deep-learning accelerator, C7x DSP, 4GB LPDDR4, 16GB eMMC, USB 3.0, Gigabit Ethernet, and BeagleBone cape compatibility.

## Quick Facts

- **Brand:** BeagleBoard
- **Board Type:** Single Board Computer
- **Price:** ~$185.62
- **Product Readiness:** developer
- **CPU:** TI TDA4VM (Dual Arm Cortex-A72 @ 2.0 GHz)
- **CPU Architecture:** ARM64
- **CPU Cores:** 2
- **RAM:** 4GB LPDDR4
- **Flash:** 16GB eMMC
- **USB:** USB 3.0 Type-C,2x USB 3.0 Type-A
- **Power:** 5V DC
- **Operating Voltage:** 3.3V
- **SD Card:** MicroSD

## Connectivity

- **Ethernet:** Gigabit Ethernet

## Open Source

- **Firmware:** Yes
- **Schematics:** Yes
- **PCB Layout:** Yes

## Compatible Firmware

Debian Linux, Armbian

## Use Cases

- ai-ml-development
- computer-vision
- autonomous-robotics
- drone-development
- media-server
- edge-computing
- prototyping

## Components

- **TDA4VM** (processor): Texas Instruments Jacinto TDA4VM with dual Arm Cortex-A72 at 2.0GHz, C7x DSP, 8 TOPS MMA deep-learning accelerator, vision processing accelerators, and six Cortex-R5F real-time MCUs — [Datasheet](https://openhardware.directory/r?to=https%3A%2F%2Fwww.ti.com%2Fproduct%2FTDA4VM&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export)
- **PowerVR Rogue 8XE GE8430** (gpu): Imagination Technologies PowerVR Rogue 8XE GPU running at up to 750MHz delivering 96 GFLOPS and 6 Gpix/sec — [Datasheet](https://openhardware.directory/r?to=https%3A%2F%2Fwww.imaginationtech.com%2Fgpu%2Frogue%2F&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export)
- **4GB LPDDR4** (memory): 4GB LPDDR4 RAM for high-bandwidth AI/ML workloads and multitasking — [Datasheet](https://openhardware.directory/r?to=https%3A%2F%2Fwww.ti.com%2Fproduct%2FTDA4VM&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export)
- **16GB eMMC** (storage): 16GB eMMC flash with high-speed interface for OS and application storage — [Datasheet](https://openhardware.directory/r?to=https%3A%2F%2Fwww.ti.com%2Fproduct%2FTDA4VM&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export)

## Protocols

- **USB 3.0**
- **PCIe (M.2 E-key)**
- **CSI (2x 4-Lane)**
- **DSI (4-Lane)**
- **DisplayPort (Mini)**
- **UART**
- **JTAG**
- **I2C**
- **SPI**

## Available Software

- **[Debian Linux]()** (open-source)
- **[Armbian Linux]()** (open-source)

## Description

## Overview

BeagleBone AI-64 is the most powerful AI open-source platform in the BeagleBoard family. Built on a proven open source Linux approach, it brings massive computing power with the Texas Instruments TDA4VM SoC featuring dual 64-bit Arm Cortex-A72 cores, C7x DSP, and deep learning accelerators up to 8 TOPS.

## Key Features

- **AI/ML Acceleration**: Deep-learning matrix multiply accelerator (MMA) up to 8 TOPS (8b) at 1.0 GHz
- **Vision Processing**: VPAC with ISP and multiple vision assist accelerators, plus DMPAC
- **DSP Power**: C7x DSP at 80 GFLOPS, 256 GOPS plus two C66x DSPs at 40 GFLOPS each
- **Real-time Control**: Six Arm Cortex-R5F MCUs at up to 1.0 GHz
- **3D Graphics**: PowerVR Rogue 8XE GE8430 GPU at 750 MHz, 96 GFLOPS
- **Cape Compatibility**: BeagleBone Black header compatibility for expansion with existing capes
- **MikroBus**: Shuttle header for hundreds of Click sensors and actuators

## Connectivity

- M.2 E-key PCIe connector for WiFi modules
- Gigabit Ethernet (RJ45)
- USB 3.0 Type-C (power + data) + 2x USB 3.0 Type-A
- Mini DisplayPort for monitors
- 2x 4-Lane CSI camera connectors
- 4-Lane DSI for flat-panel displays
- 2x UART debug + JTAG

## Software Support

- Debian Linux (XFCE and Minimal images)
- Armbian Linux
- Full open-source toolchain

## Where to Buy

- [DigiKey](https://openhardware.directory/r?to=https%3A%2F%2Fwww.digikey.com%2Fen%2Fproducts%2Fdetail%2Fbeagleboard-by-seeed-studio%2F102110646%2F15929655&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export) — $185.62
- [Mouser](https://openhardware.directory/r?to=https%3A%2F%2Fwww.mouser.com%2FProductDetail%2FBeagleBoard-by-Seeed-Studio%2F102110646%3Fqs%3DZnm5pLBrcAKSPlLdpqfmWA%253D%253D&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export) — $185.62
- [Element14](https://openhardware.directory/r?to=https%3A%2F%2Freferral.element14.com%2FOrderCodeView%3Ffsku%3D3923739%26nsku%3D26AK8283&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export) — $185.62

## Resources

- [product](https://openhardware.directory/r?to=https%3A%2F%2Fwww.beagleboard.org%2Fboards%2Fbeaglebone-ai-64&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export)
- [github](https://openhardware.directory/r?to=https%3A%2F%2Fopenbeagle.org%2Fbeagleboard%2Fbeaglebone-ai-64&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export)
- [wiki](https://openhardware.directory/r?to=https%3A%2F%2Fdocs.beagleboard.org%2Flatest%2Fboards%2Fbeaglebone%2Fai-64%2F&ref=openhardware.directory&product=beaglebone-ai-64&source=md-export)

## Tags

`beagleboard`, `ai`, `machine-learning`, `arm-cortex-a72`, `tda4vm`, `single-board-computer`, `linux`, `open-hardware`

## Images

![BeagleBone AI-64](https://www.beagleboard.org/app/uploads/2022/04/BeagleBone_AI_64.webp)

---
[View full device page](https://openhardware.directory/devices/beaglebone-ai-64)