What is OpenGradient (OPG)

I. Project Introduction

1. What Is Opengradient?

OpenGradient is a decentralized, high-performance, and verifiable computing infrastructure specifically built for AI. It aims to achieve user-owned, portable intelligence, enabling builders to use auditable open-source AI models, host models on-chain, run secure inference, and deploy agents and applications. By combining the security of blockchain with the optimization capabilities of AI, OpenGradient addresses the "black box" problem of centralized AI and the limitations of blockchain in complex computation, realizing a truly sovereign AI infrastructure.

2. How Does Opengradient Operate?

OpenGradient utilizes a unique hybrid AI computing architecture. This architecture is designed specifically for the characteristics of AI workloads, separating execution from verification.

Inference Nodes: Use GPU and TEE (Trusted Execution Environment) nodes to execute model inference and return results directly to users. This supports a "fast path" with latency approaching that of centralized APIs.

Full Nodes: Responsible for consensus, verifying cryptographic proofs (TEE proofs, ZKML proofs, or signatures) using the CometBFT mechanism, settling payments, and recording the ledger.

Data Nodes: Securely obtain external data through TEEs and provide corresponding proofs.

Storage Layer: Relies on decentralized storage (such as Walrus) to handle model files and large proofs, recording only blob IDs on-chain to ensure maximum efficiency.

3. Who Created Opengradient?

OpenGradient was co-founded by Matthew Wang (CEO & Co-Founder) and Adam Balogh (CTO & Co-Founder).

Matthew Wang has a background in quantitative research and software engineering, with previous experience at Two Sigma, Google, Facebook, and NASA.

Adam Balogh previously served as the Tech Lead for the Palantir Artificial Intelligence Platform and possesses engineering experience from companies such as Google and Amazon.

The team brings together global talent from top-tier institutions like Google, Meta, and Palantir, and is headquartered in New York. The project is supported by prominent institutions including a16z crypto and Coinbase Ventures, along with strategic endorsements from industry leaders such as Balaji Srinivasan, Illia Polosukhin, and Sandeep Nailwal.

4. OpenGradient Tokenomics

$OPG is the economic layer token of the OpenGradient network, with a fixed total supply of 1 billion tokens. $OPG is used for paying inference fees, model monetization, staking for security, application access, and governance.

Token Allocation and Vesting Schedule:

Ecosystem (40%): 10% at TGE, then the remaining 30% linearly over 60 months.

Foundation (15%): 33.33% at TGE, then the remaining linearly over 48 months.

Core Contributors (15%): 12-month cliff, then 36-month linear vesting.

Investors and Advisors (10%): 12-month cliff, then 36-month linear vesting.

Staking Rewards (10%): Linearly over 96 months.

Liquidity and Launch (6%): 100% unlocked at TGE.

Airdrop (4%): 100% unlocked at TGE.

5. Timeline & Key Milestones

October 6, 2024: Official launch of OpenGradient, introducing the Model Hub and OG SDK, and announcing the roadmap.

October 2024: DevNet goes live; developers begin building on-chain AI applications.

2025: Testnet launch; completion of the Seed funding round totaling approximately $9.5 million.​

II. Token Information

III. Related Links