Skip to main content

Subcloud Use Cases

Subclouds provide flexibility in how inference workloads are routed, hosted, and managed within the Resultity network.
They enable fine-grained control over privacy, performance, model availability, and regulatory compliance.

Corporate Clusters​

  • Enterprises can create private Subclouds with selected nodes (e.g., self-hosted or trusted vendors);
  • Ensures full control over where and how jobs are processed;
  • May include internal models, business logic, or restricted data access.

Custom or Fine-Tuned Models​

  • Subclouds allow deployment of fine-tuned or proprietary models separate from the global pool;
  • Ideal for specialized domains such as legal, medical, finance, or local languages;
  • Jobs targeting these models are routed only to eligible nodes.

SLA Pools​

  • Nodes that meet latency, availability, or GPU class guarantees can form SLA-compliant Subclouds;
  • Users or organizations requiring stable performance can route requests to these pools;
  • SLA enforcement and monitoring is handled by the central layer.

Regional Compliance​

  • Subclouds can be created based on node geography (e.g., EU-only, US-only, Indonesia-only);
  • Enables jurisdiction-specific compliance (e.g., GDPR, data residency laws);
  • May use regional gateways and routing optimizations.

Thematic or Model-Type Subclouds​

  • Dedicated clusters for certain types of models or use cases:
    • Multimodal Subclouds (e.g., image + audio models);
    • Open-weights only (no proprietary models);
    • Healthcare-only clusters (subject to enhanced trust levels).

Partner and Community Subclouds​

  • Resultity partners (e.g., universities, research groups, cloud providers) can launch branded Subclouds;
  • May include community governance, public dashboards, or co-incentives;
  • Ideal for growing decentralized capacity with trust anchors.

Subclouds offer a modular, extensible way to structure the decentralized network without fragmenting its core.
They allow specialization, compliance, and performance guarantees — while maintaining interoperability.