Skip to main content

Hyperbolic

Web3 infrastructure.

About Hyperbolic​

Hyperbolic operates as an open-access AI cloud platform that aggregates decentralized global GPU resources to provide scalable, affordable AI infrastructure and services. Founded by researchers from UC Berkeley and the University of Washington, its mission is to democratize AI by solving compute scarcity, verification, and privacy challenges in decentralized environments. Key offerings include a decentralized GPU marketplace (costing up to 75% less than traditional clouds) and AI inference services supporting 38+ models, including open-source and proprietary options. In Web3, Hyperbolic bridges blockchain and AI by enabling verifiable, trustless AI workflows through its Proof-of-Sampling protocol and decentralized orchestration layer.

Technology​

Hyperbolic’s stack combines:

  • Decentralized GPU Network: Aggregates idle GPUs globally into a marketplace, managed by Hyper-dOS (decentralized operating system).
  • Verification Layer: Uses blockchain and Proof-of-Sampling (PoSP) to ensure outputs are generated by correct models and hardware.
  • Privacy Layer: Implements confidential computation to protect user data during third-party node processing.
  • AI Services: Inference-as-a-service with low-latency processing of 1B+ tokens daily, compatible with frameworks like HuggingFace and TensorFlow.

This solves GPU scarcity, reduces AI compute costs, and guarantees integrity for decentralized applications.

Key Features​

  • Cost-Efficient Compute: 75% savings vs. centralized clouds via idle GPU utilization.
  • Multi-Model Inference: Supports Llama 3, Mistral, and custom models with seamless API integration.
  • Trustless Verification: PoSP cryptographically validates AI outputs.
  • Hybrid Ecosystem: Serves both Web2 (Stanford, NYU) and Web3 (Virtuals, Near) users.
  • Fault-Tolerant Design: Autonomous recovery from network disruptions.
  • Monetization: GPU owners earn rewards by renting idle resources.
  • Developer Tools: SDKs for deploying AI agents, smart contracts, and multi-chain apps.

Integration with Eliza​

Hyperbolic powers ElizaOS’s AI agents via a dedicated plugin, enabling:

  1. Compute Backend: ElizaOS agents use Hyperbolic’s GPU network for inference, reducing latency and costs.
  2. Verifiable Agents: PoSP ensures agent decisions are transparent and auditable on-chain.
  3. Plugin Configuration: Developers inject Hyperbolic’s API key and model preferences into ElizaOS environment variables, allowing direct access to 10,000+ GPUs.
  4. Use Cases: Autonomous agents for DeFi, cross-chain arbitrage, and NFT analytics leverage Hyperbolic’s scalable infrastructure.

The integration is documented in Hyperbolic’s GitHub repositories, with templates for deploying ElizaOS agents on decentralized compute.

Recent Developments​

  • $12M Series A Raise (Dec 2024): Co-led by Variant and Polychain, totaling $20M to date.
  • GPU Marketplace Launch: Opened to all users after waitlist phase, now hosting 10,000 GPUs.
  • e/acc Initiative: Experimental Acceleration program for AI agent R&D, processing 59M requests in Q1 2025.
  • Blockchain Development: Building a dedicated chain for GPU resource allocation and tokenomics.

Market Position​

Hyperbolic leads in decentralized AI infra, competing with Render and Flux but differentiated by its verification layer and hybrid Web2/Web3 adoption. Key partnerships include:

  • ai16z: Powers ElizaOS’s multi-agent framework.
  • Virtuals Protocol: Provides compute for AI agent governance.
  • Near Protocol: Collaborates on AI-driven smart contracts.

Metrics:

  • 55,000+ developers building on the platform.
  • Processes 1B+ inference tokens daily.
  • Powers 90% of top crypto-AI projects.

Website: hyperbolic.xyz
GitHub: github.com/hyperbolic
Documentation: docs.hyperbolic.xyz
Discord: Discord Server
ElizaOS Integration Guide: Blog Post

This profile reflects Hyperbolic’s role as a critical infrastructure layer for Web3 AI, combining decentralized compute with rigorous verification to enable trustless, scalable applications.