Remote AI computing, built from the ground up.
Orca Systems is developing a hardware and software platform designed to make powerful AI computing accessible remotely, from anywhere in the world.
Rather than requiring users to own expensive GPUs, Orca aims to connect them to remote computing hardware over the internet for AI inference, development, and other GPU-intensive workloads.
Orca Systems brings together three core components:
- Hardware — Compact GPU computing devices built using off-the-shelf accelerators, custom electronics, networking, and cooling.
- Software — A platform for connecting users to computing devices, managing workloads, and executing tasks remotely.
- Distributed computing — A network of Orca devices intended to make computing resources accessible across different physical locations.
The long-term goal is to bring these components together into a unified platform for remote AI computing.
Development begins with existing hardware.
The first prototype aims to transform a GPU-equipped PC into a remotely accessible AI compute node. This provides a practical foundation for developing and testing the software before moving towards dedicated Orca hardware.
Initial development focuses on the foundations of remote compute access, workload execution, and device management.
Orca Systems aims to make AI computing more accessible, scalable, and less dependent on traditional cloud infrastructure.
The longer-term vision is a platform combining purpose-built hardware, distributed compute resources, and a unified software experience.
Orca's advantages over existing GPU cloud services have yet to be demonstrated and will be evaluated as the technology develops.
- Initial prototype — Establish a remotely accessible AI compute node using existing PC hardware.
- Remote execution — Develop the core software for accessing and executing workloads on remote hardware.
- Compute management — Explore device availability, workload management, and resource allocation.
- Multi-node support — Investigate coordinating compute resources across multiple devices.
- Dedicated hardware — Explore compact, purpose-built Orca computing devices.
- Distributed network — Work towards a unified platform for accessing distributed AI compute resources.
Roadmap items represent development goals, not completed features.
Orca Systems is in its early development stages. Architecture, features, hardware designs, and priorities may change as the prototype evolves.
Orca Systems' source code is publicly viewable, but it is proprietary and source-available, not open-source software.
No permission is granted to use, modify, reproduce, or redistribute the code except where expressly permitted by the applicable licence or by written authorisation from Orca Systems.
See LICENSE for the applicable terms. For licensing enquiries, visit orcaservers.com.
- Website: https://orcaservers.com
- Source code: Available through the relevant Orca Systems GitHub repositories
© Orca Systems. All rights reserved.
