Skip to main content

Understanding Orbits

What are Orbits?​

Orbits are specialized AI networks within Orbinum where miners compete to provide the highest quality inference services for specific AI domains. Each Orbit operates as an independent marketplace with its own quality standards, miner rankings, and emission distribution.

Think of Orbits as specialized guilds where AI miners compete to demonstrate expertise in their chosen domain--whether that's natural language processing, computer vision, or predictive analytics.

This page provides a high-level overview of Orbits. For detailed guides on creating, managing, and participating in Orbits, see the Orbits section.

Orbit Types​

Initial Orbits (Public Testnet)​

During the public testnet phase, the Orbinum team will launch two foundational Orbits to bootstrap the ecosystem:

  • NLP Orbit (Text): Text generation, summarization, translation, and question-answering.
  • Code Orbit (Developer Tools): Code completion, security auditing, and automated refactoring.

These initial Orbits serve as reference implementations, demonstrating quality standards, evaluation metrics, and best practices for future community-created Orbits.

Future Orbit Types​

After the testnet phase, the network will expand through governance-funded bounties that incentivize the community to build additional Orbits, such as:

  • Vision Orbit (Images & Video): Classification, object detection, image/video generation and analysis.
  • Prediction Orbit (Forecasting): Time-series forecasting, anomaly detection, and market predictions.
  • Audio Orbit (Speech & Sound): Speech-to-text, text-to-speech, speaker ID, and audio classification.
  • Embeddings Orbit (Vectors): Semantic embeddings for search, similarity, clustering, and recommendations.

Each Orbit is an independent marketplace focused on a specific class of AI tasks; miners compete on quality and availability to earn request volume and rewards.

How Orbits Work​

Orbits are the runtime marketplaces where agents send tasks and miners provide inference. The basic flow:

  • Agent submits a request to the network.
  • The Orbit registry selects candidate miners based on Orbit type and rank.
  • A miner executes the task and returns a result.
  • Validators score the result and the registry updates miner rankings.

Key mechanics​

  • Miner registration: stake 1,000 $ON per Orbit to register.
  • Quality & ranking: validators evaluate outputs; ranking drives traffic.
  • Routing: higher-ranked miners receive most requests; lower-ranked get fewer.
  • Rewards: miners earn from user fees and block emissions, split by quality share.

Economic Model​

Orbits generate value through block emissions and user fees, which are distributed among miners, validators, Orbit owners, and stakers.

Key Distribution:

  • Block Emissions: 80% to miners, 20% to validators
  • User Fees: 2% to Orbit owner, then 65% to miner, 18% to validators, 15% to stakers

Participant Requirements:

  • Orbit Owners: Stake 1,000 $ON (locked, not burned)
  • Miners: Stake 1,000 $ON per Orbit
  • Validators: Stake required to participate
  • Stakers: Earn passive real yield

For a complete breakdown of revenue streams, costs, and ROI calculations, see Orbit Economics.

Creating Custom Orbits​

Anyone can create an Orbit by staking 1,000 $ON. New Orbits enter a mandatory 2-week testing period where validators evaluate their quality, stability, and network compatibility. After testing, Orbits are automatically activated or rejected based on evaluation results.

Community Bounty Program (Phase 2): After the testnet phase and the transition to Open Governance (Phase 2), the DAO will launch governance-funded bounties to incentivize community developers to build specialized Orbits for emerging AI domains.

Key Requirements:

  • Stake 1,000 $ON (locked, not burned)
  • Pass 2-week quality evaluation
  • Pay validation fees during testing
  • Maintain minimum quality standards

For a complete step-by-step guide, technical requirements, and best practices, see Create an Orbit.

Governance & Lifecycle​

Orbit creators can choose between two governance models:

Open Governance:

  • Parameter changes require community discussion and DAO voting
  • Ideal for protocol-level and community-driven Orbits

Closed Governance:

  • Creators retain full control over parameters
  • Faster iteration cycles for specialized or experimental Orbits
  • DAO retains veto power for critical security issues

Lifecycle Stages: Registration → 2-Week Testing → Quality Evaluation → Activation/Rejection → Operation → Deactivation

All Orbits must maintain minimum quality standards. The DAO can deactivate any Orbit that poses network security risks.

For detailed governance mechanics, dispute resolution, and Orbit owner rights, see Orbit Governance.

Quality & Evaluation​

Validators evaluate miner outputs using domain-specific metrics (accuracy, latency, availability). These scores determine miner rankings, which drive request routing and reward distribution.

For in-depth details on quality evaluation mechanisms, see Quality Evaluation.

Participation Strategies​

For Miners:

  • Specialization: Focus on mastering one Orbit
  • Diversification: Participate in multiple Orbits to spread risk

For Users:

  • Choose Orbits by quality scores, latency, and pricing
  • Monitor performance via the Orbit Explorer dashboard

Future Orbit Specializations​

Potential domains for community-built Orbits include multimodal AI, advanced reasoning, robotics control, gaming AI, and creative generation.

Next Steps​