Community FAQ

June 25, 2025 marks a major milestone for the Theta Network with the launch of the Theta EdgeCloud Hybrid Beta — a decentralized platform built to power high-performance computing at scale. With a hybrid architecture that leverages both traditional cloud GPUs and community-run edge nodes, EdgeCloud is designed to support AI inference, video processing, 3D rendering, and more.
Whether you’re contributing GPU power or submitting workloads to be completed on EdgeCloud, this FAQ will help you get started.

What Is Theta EdgeCloud?

Theta EdgeCloud is a decentralized GPU compute marketplace. It connects those who need access to GPU resources with those who have unused GPU capacity. Contributors can earn rewards in TFuel, while developers and organizations gain access to scalable, cost-efficient infrastructure.

For Node Operators: Contribute Your GPU and Earn TFuel

What is the EdgeCloud Client?

The EdgeCloud Client is a lightweight software package that lets you contribute your GPU power to Theta EdgeCloud. Once installed, your machine becomes part of a global infrastructure capable of running containerized jobs.

System Requirements

Recommended hardware:

GPU: NVIDIA 3090, 4090, A100, H100 or similar (8GB vRAM minimum)
CPU: Intel or AMD with 4 or more cores
RAM: 16GB or more
Disk: At least 256GB of free storage
Internet: 100 Mbps upload and download
OS: Ubuntu 22.04+ (recommended), Windows 10/11 (Alpha support)

Machines meeting the above specs will be much more likely to get high-reward jobs. Moreover, clients who are used to cloud services like AWS or GCP may come with high expectations, so it’s a good idea to aim for consistent uptime. When being used, the GPU will likely be running near full capacity, so make sure your internet connection, power supply, and cooling systems are in good shape.

How Are Rewards Calculated?

Rewards are calculated in USD and converted to TFuel
Payouts are made monthly between the 1st and 5th of each month
Node operators can set their own hourly GPU rental rate
The marketplace model allows demand-side users to select nodes based on price and performance

Getting Started

To install and run the EdgeCloud client, follow the official documentation:

EdgeCloud Client Installation Guide
Command-line (CLI) and RPC API references are also available for advanced users

For Job Submitters: Run High-Performance Workloads on EdgeCloud

What Types of Jobs Can I Run?

Theta EdgeCloud supports any containerized job requiring GPU acceleration, including:

AI inference and generative AI models
Text-to-image and chatbot applications
Video encoding and transcoding
3D rendering pipelines
Financial modeling and scientific simulations

Jobs are routed to the most optimal mix of edge and cloud GPUs for efficient and cost-effective processing.

Submitting a Job

You can submit containerized jobs on Theta EdgeCloud to be completed on community-run GPUs using this guide. There are three initial ways to leverage community Edge Nodes:

Use Community Nodes as GPU Machines with SSH Access
Dedicated Model Deployment Onto Community Nodes
On-demand Model APIs Leveraging Community Nodes

Can I Select Specific GPUs?

Yes. Users can filter available nodes based on:

GPU type (e.g., 4090 vs A100)
Hourly rate
Performance characteristics

This flexibility helps users balance cost and speed based on their project requirements.

What Happens If a Node Goes Offline?

EdgeCloud supports:

Automatic failover
Job containerization
Intelligent reallocation of workloads

If a node goes offline mid-task, the system will reroute the job to another available GPU to ensure completion without manual intervention.

Developer Features in Beta

Available at launch:

API access for job submission
Intelligent job routing across edge and cloud
Node performance monitoring

Coming soon:

Job opt-out and preference settings for node operators
Enhanced marketplace insights

Who Should Use Theta EdgeCloud?

Theta EdgeCloud is designed for:

Developers building GenAI tools and AI agents
Video professionals handling large-scale transcoding
Researchers running scientific simulations
GPU owners looking to earn passive income from idle hardware