top of page
Your Software in Space

Run your own AI/ML software models in space

Exo-Space offers the ability to run a customers artificial intelligence/machine learning (AI/ML) software models on Earth Observation satellites without the customer having to launch their own. Software models are hosted on Exo-Space’s FeatherEdge payload, comprising of its own dedicated CPU, GPU and AI accelerator, capable of performing 4 trillion operations per second. The FeatherEdge payload receives images in real time, directly from the satellites onboard camera and runs the raw images through the customers AI/ML models. Processed data is sent directly to the customer after downlink through an easy to use API. 


Customers have the flexibility to specify desired areas of interested for imaging, upload and update the models at any time, and receive the processed data according to their schedule.

What We Offer

Run your own TensorFlow Lite models onboard an Earth observation satellite

Process satellite imagery directly from the onboard imager and get access to the images

Service includes mission operations and data communication assuming 3 ground station passes per day

Access to a development environment for ground testing

Unlock the potential of AI in Space


Computing Hardware

Powerful machine vision processor and AI accelerator.

Ability to upload multiple models at once and change mission needs at any time.


Simple monthly fee structure for hosting software models.

Reduced Cost

bottom of page