Artificial Intelligence Radio - Transceiver (AIR-T)

by Deepwave Digital

A high-performance SDR seamlessly integrated with state-of-the-art deep learning hardware

Crowdfunding Page History

This page is an archive of the original crowdfunding campaign for this project. It may not be up-to-date with the latest updates and product availability. Return to the current project page.

$165,515 raised

of $20,000 goal

827% Funded! Order Below

Pledge Now


Support Us!

Thanks for helping make the AIR-T a reality! You will be rewarded with good karma and regular backer updates.



The AIR-T is the first platform that enables out-of-the-box machine learning wireless systems. Simply port current GNURadio or code directly in Python to access the radio frequency spectrum. Includes, AIR-T board, four MCX-to-SMA cables, getting started tutorial, and power supply.


MCX-to-SMA Cables (6 Pack)

A set of six MCX-to-SMA cables allows for easy access to all the AIR-T MCX connections simultaneously. Each cable is 6 inches (15.24 cm) long with a male MCX right-angle connector and a female SMA (bulkhead style) connector.


Recent Updates

As Featured In


"The basic idea behind the AIR-T is to combine a 2x2 MIMO SDR transceiver with a NVIDIA Jetson TX2 GPU that can be used to run artificial intelligence (AI) software fast."


"Con una potente GPU programable de NVIDIA, este transceptor de radio con inteligencia artificial facilita la creación de aplicaciones autónomas."


"A whole lot beefier than other [SDRs] I've seen around... if you're a radio-head, then this one would interest you."

CNXSoft - Embedded Systems News

"The system can be used as a highly parallel SDR, data recorder, or inference engine for deep learning algorithms, while the GPU enables SDR applications to process bandwidths greater than 200 MHz in real-time."

Geeky Gadgets

"Created for new engineers and enthusiasts with little wireless experience to advanced engineers and researchers who develop low-cost AI, deep learning, and high-performance wireless systems."

"The AIR-T is designed for researchers who want to apply the deep learning powers of the Jetson TX2’s 256-core Pascal GPU and its CUDA libraries to the SDR capabilities provided by the Artix 7 and AD9371 transceiver."

Military Embedded Systems

"Deepwave Digital's aim is to enable the next-generation of RF systems to utilize deep learning by combining high performance computing (HPC) with a SDR in a single embedded platform."

NVIDIA has officially backed the AIR-T campaign. Read the full update here.

Why build tomorrow’s tech with yesterday’s signal processing tools? The Artificial Intelligence Radio - Transceiver (AIR-T) is a fully integrated, single-board, artificial intelligence equipped, software defined radio platform with continuous frequency coverage from 300 MHz to 6 GHz. Designed for new engineers with little wireless experience to advanced engineers and researchers who develop low-cost AI, deep learning, and high-performance wireless systems, AIR-T combines the AD9371 RFIC transceiver providing up to 2 x 2 MIMO of 100 MHz of receiving bandwidth, 100 MHz of transmitting bandwidth in an open and reprogrammable Xilinx 7 FPGA, with fast USB 3.0 connectivity.

The AIR-T has custom and open Ubuntu software and custom FPGA blocks interfacing with GNU Radio, allowing you to immediately begin developing without having to make changes to existing code. With 256 NVIDIA cores, you can develop and deploy your AI application on hardware without having to code CUDA or VHDL. Freed from the limited compute power of a single CPU, with AIR-T, you can get right to work pushing your telecom, defense, or wireless systems to the limit of what’s possible.

Three Integrated Processors

Deepwave’s AIR-T is the first software-defined radio (SDR) with embedded high-performance computing. The AIR-T lowers the price and performance barriers to autonomous signal identification, interference mitigation, and much more. The AIR-T allows for a fully autonomous SDR by giving the AI engine complete control over the hardware. Whether your background is in electrical engineering, applied physics, or a related field, AIR-T will open up new possibilities for your work. It does this by uniquely integrating three digital processors that provide the functionality needed for any signal processing application:

  • FPGA for strict real-time operations
  • GPU for highly parallel processing
  • CPU for control, I/O, DSP, and software applications

This unique combination brings you the worlds of high-performance computing (HPC), artificial intelligence, deep learning, and advanced graphics and rendering on one embedded platform. The system has the versatility to function as a highly parallel SDR, data recorder, or inference engine for deep learning algorithms. The embedded GPU allows your SDR applications to process bandwidths greater than 200 MHz in real-time.

Features and Specifications

  • Software-defined Radio
    • Analog Devices 9371 2x2 MIMO transceiver
      • 2 x RX channels (100 MHz each)
      • 2 x TX channels (100 MHz each)
      • Auxiliary RX channels*
        • Observation
        • Sniffer
    • NVIDIA Jetson TX2 for processing
      • 256 NVIDIA core integrated GPU
      • 6 CPU cores
        • Dual-core NVIDIA Denver2
        • Quad-core ARM Cortex-A57
      • 8 GB CPU/GPU shared memory
    • Xilinx Artix-7 FPGA
      • 75k logic cells
  • Connectivity
    • USB 2.0 and 3.0
    • 1 Gig Ethernet
    • HDMI display port
    • High-speed digital I/O (GPIO/UART)
    • 1 PPS & 10 MHz reference input
    • External LO input
  • Storage
    • 32 GB of on-board storage
    • External USB
    • SATA
    • SD card
  • Software
    • SD card w/ Ubuntu 16.04
    • SoapySDR - Port existing GNU Radio applications within minutes
    • Custom GNU Radio blocks to easily deploy your neural network or high-performance applications
    • Build you own custom software using our hardware drivers
    • Use Python or C++ to easily program the GPU
    • No VHDL/Verilog. No need to learn how to program an FPGA (but you can if you want to!)
  • Physical
    • Form factor: Mini-ITX
    • Dimensions: 170 mm × 170 mm x 35 mm (6.7” × 6.7” x 1.4”)
    • Weight: 0.35 kg (0.8 lbs)
    • Power: 22 Watts (14 Watts with reduced GPU cores)

* Can use either Observation or Sniffer at one time; utilizes one of the RX channels

AIR-T functional mechanical drawing

What’s In The Box?

  • AIR-T board
  • 4 MCX-to-SMA cables
  • SD card w/ customized Ubuntu operating system and software
  • Getting started tutorial
  • Power supply
  • Optional items:
    • Enclosure
    • Additional MCX-to-SMA cables


AIR-T Ettus E310 Ettus N310 LimeNET Mini Epiq Maveriq
GPU for Signal Processing 256 Core NVIDA Jetson - - - -
Deep Learning Capable TensorFlow, Caffe, Keras - - - -
CPU Cores 6 (ARM A57, Denver2) 2 (ARM A9) 2 (ARM A9) 2 (Intel i7-7500U) 4 (Intel Atom)
RAM (GB) 8 1 1 32 8
Internal Storage (GB) 32 - - 512 Up to 1000
Tx Bandwidth > 60 MHz 100 MHz - 100 MHz 61.4 MHz -
Rx Bandwidth > 60 MHz 100 MHz - 100 MHz 61.4 MHz -
Max Bandwidth for Onboard Processing (MHz) >200 10 Not Published Not Published Not Published
USB 3.0 1 - - 2 -
SATA 1 - - 1 (Internal Storage) 1 (Internal Storage
1 Gb Ethernet 1 1 1 1 1
Wi-Fi 1 - - 1 -
Bluetooth 1 - - 1 -
Display Out HDMI (4K) - - HDMI (4K) -
Max Power Consumption (W) 22 (estimated) 6 80 Not Published 14
Price $6,299 $5,670 $2,982 $10,000 $2,599 Unavailable

Overview Video

Use Cases


With the AIR-T, you can use deep learning to maximize applications from Wi-Fi to OpenBTS. Pairing a GPU directly with an RF front-end means you don’t have to purchase an additional computer or server for processing. Just power on the AIR-T, plug in a keyboard, mouse, and monitor and you’re ready to go. Use GNU Radio blocks to quickly develop and deploy your current or new wireless system or, if you need more control, talk directly with the drivers using Python or C++. And for you superusers, the AIR-T is an open platform, so you can program the FPGA and GPU directly.

Satellite Communications

Communicating past Pluto is hard. With the power of a single-board SDR with an embedded GPU, the AIR-T can prove your concepts before you launch them into space. AIR-T lets you reduce development time and costs by adding deep learning to your satellite communication system. With the ability to program in Python and rapidly port existing code from GNU Radio, you can accelerate your existing applications within minutes. Yet, you can do a LOT more with the AIR-T. We are committed to having an open architecture. Meaning, you can program in Python, control the drivers with your a custom software, or program the FPGA directly. We anticipate customers developing at every level.

Ground Communications

There is a seemingly endless number of terrestrial communication systems, with more being developed every day. From high-power, high-frequency voice communications to 60 GHz millimeter wave digital technology, there are significant challenges in every band. As spectral density becomes more congested, we are nearing the end of the amount of information that can be passed over wireless systems using the same spectrum. AI can be used in order to maximize these resources. The AIR-T is well-positioned to easily and quickly help you prototype and deploy your wireless AI solution. From 300 MHz to 6 GHz the AIR-T covers the majority of commercial communication bands.

Video/Image/Audio Recognition

While the AIR-T was designed for wireless development, it can process any type of data. NVIDIA is a graphics processing powerhouse and their products, including the Jetson TX2, are known for their high-performance when it comes to video. With the AIR-T, you can combine the traditional uses of image and video processing with radio frequency. With USB 3.0 & 2.0, Gigabit Ethernet, and high-speed IO, there are many ways to bring data in and out of the AIR-T. You can attach additional sensors and allow the AIR-T to fuse the data together.

The AIR-T lets you demodulate a signal and apply deep learning to the resulting image, video, or audio data in one integrated platform. For example, you used to need multiple devices to directly receive a signal that contains audio and then perform speech recognition. The AIR-T integrates this hardware into one easy-to-use package. From speech recognition to digital signal processing, the integrated NVIDIA GPU provides the horsepower needed for your cutting edge application.

Software Flow

Stretch Goals

Our minimum goal is to sell four AIR-T units. If we sell ten or more units, we’ll publish to all AIR-T buyers a “Getting Started Guide: End-to-end Machine Learning,” showing you how to develop a data recording and labeling scheme for Automatic Dependent Survallience-Broadcast (ADS-B) along with functional TensorFlow code to train a convolutional neural network to detect the ADS-B signal on your own using the AIR-T. If we sell 20 or more units, we’ll include a metal enclosure to keep your AIR-T safe and secure.

Manufacturing Plan

Nothing great is created without a team. At Deepwave, we believe in partnering with the best in the industry. Our engineering team and contract manufacturer have created boards that have been launched into space and been to deepest parts of the ocean.

Before making our selection, we met with numerous contract manufacturers, toured their facilities, and understood their commitment to quality. We found a CM that is experienced in creating transceivers, can achieve extremely high tolerances, routinely works with small businesses and startups, and, most importantly, has a commitment to quality. This CM has deep experience well-matched to meet our short-term needs and long-term goals. They have developed a manufacturing process to handle prototype runs within a few days as well as large production runs in the tens and hundreds of thousands.

Deepwave is intimately involved with the manufacturing process at every step to ensure the highest quality. We believe we have selected a partner who can achieve our standards for quality and timeliness.

All we need are the first few backers to make this product a reality. Once an AIR-T is manufactured, it will be functionally tested by a member of the Deepwave staff, packaged up, and shipped out.

Shipping & Fulfillment

We would love to hand-deliver each one, but that’s a little impractical. However, a member of the Deepwave staff will be shipping the products. One of the benefits of selecting a manufacturer close to our headquarters is that we are able to easily visit and meet with their staff. Here is the process:

  • AIR-T Manufactured. As each board is finished, it will be tested by a Deepwave engineer to ensure it meets our performance specifications.
  • Packaged Up. Once the board passes our quality test, we will package it up. In working with our contract manufacturer we are able to package it right at their facility.
  • Shipped. Once it is packaged up, we will ship it directly to you. This reduces the downtime and unnecessary time to bring it to a third-party shipping facility. (Future units may be shipped from another location, but our initial units will be direct.)

UPS is the carrier and we ship internationally, too.

Risks and Challenges

There are always risks in any endeavor. Building an industry-leading, high-performance, edge-computing, software-defined radio platform (whoa… that’s a mouthful) comes with a few risks we feel you should know about. The most important thing to know is that we have addressed them.

Experience. The risk we identified from the first moments of our company is in the experience of the team. The answer to that is “YES, we have the experience.” Our team consists entirely of engineers (and some advisers) who have all designed, built, deployed, and supported similar systems before. We started this company to design a system that we wanted to use. Our average team member has 12 years of experience directly in this industry. From research scientists and RF engineers to full-stack wireless developers, we have the depth and breadth of experience to bring this product to you.

Physical Defect. The second risk is if something is wrong with how the circuitry is run, or if there is a physical problem. This could be interference from another circuit or something run incorrectly. We have done a few things to alleviate this. The first is that we have had multiple external reviews, meaning multiple independent engineers not from our team, who are experts in the field, have performed a detailed review. This is common for this type of work and helps identify potential problems without having a day-to-day influence on the design. Just to make sure, we are also making a small number of boards before we go to full-rate production.

Manufacturing. The third risk is contract manufacturing. When a third party builds your hardware, there is a huge amount of reliance on those partners. All electronics manufacturers are not equal and their commitment to quality is often wildly different. Also, the teams need to work together. You need to be able to grow together and meet long-term needs. The last thing we want is to have to switch manufacturers for the same product. This introduces unnecessary risk, additional tooling costs, and the possibility of an error. We have selected a manufacturer that can handle everything from a small number of boards up to hundreds of thousands of boards.

Sourcing Parts. The electronics industry is facing huge problems right now trying to accurately match supply with demand. Some components are difficult to purchase because of long lead times. We’ve taken many steps to avoid this problem such as designing out some parts and finding alternatives to others.

What we have presented in this section is a list of the top risks. As in any business, there are many other smaller decisions that happen day-to-day that would constitute risks, however, these are simply part of running any company. If something changes fundamentally impacting our promise to deliver a quality product, you will hear from us.

Get in Touch

If you have any questions or simply want to find out more about AIR-T or the Deepwave team, please don’t hesitate to get in touch.


Deepwave Digital

We are a dedicated team of radio frequency and wireless industry experts with over fifty years of combined experience providing digital signal processing solutions to the commercial and defense industries. We understand both the technical capabilities as well as the limitations of these technologies, leading us to develop novel hardware and software solutions to combine the fields of artificial intelligence and signal processing.

John Ferguson

Steven Kline

Peter Witkowski

Steve Lisi

Subscribe to the Crowd Supply newsletter, highlighting the latest creators and projects: