2 million users and 10k agents on the AI blockchain

Over 2 million registered users and 10,000 agents already published. ChainOpera brings decentralized AI from the experimental phase to a concrete implementation on blockchain, with incentives, traceability, and integrated verifications, as described in the official ChainOpera whitepaper and in industry analyses published on The Block.

The project, described by co-founder Salman Avestimehr – professor at USC and recipient of the U.S. Presidential Early Career Award for Scientists and Engineers (PECASE), as well as IEEE Fellow – combines decentralized federated learning and distributed ledgers to build an operational and interoperable ecosystem.

In this context, the stated goal is to return ownership and control to communities of developers, computing providers, and end users.

According to the public data available in the whitepaper and official communications, updated as of September 30, 2025, the platform reports over 2 million registered users and approximately 10,000 published agents.

Industry analysts note that the completion of the $17 million seed round in December 2024 accelerated the development of the L1 and OS components. Cross-checks of the technical pages reveal metrics on PoI functionalities and the layered structure that confirm the approach described in the project document.

Origins, team, and vision

ChainOpera was born from the experience gained with FedML, an open-source library launched in 2019 for decentralized federated learning. The project was co-founded by Salman Avestimehr and Chaoyang Aiden He, who initiated FedML during his PhD FedML ChainOpera.

The team includes academic profiles and engineers from UC Berkeley, Stanford, USC, MIT, and companies like Amazon, Google, and Microsoft.

The long-term vision converges towards a community-owned and governed “Crypto AGI,” a topic addressed in interviews and debates within the sector The Block. The $17M seed round was announced in December 2024 and saw the participation of institutional investors supporting the development of the L1 infrastructure.

What does “decentralized AI” on blockchain mean

A decentralized AI distributes training, execution, and verification of agents across a network of nodes without central control.

The blockchain, in this context, records contributions, certifies results, and manages rewards according to transparent rules, as explained in the protocol specifications and the official whitepaper.

In practice, each actor – whether it be data, models, GPUs, or agents – obtains a verifiable trace of their activities. The result is an increase in transparency, improved auditability, and incentives aligned with quality.

Stack and numbers: the platform in four levels

According to ChainOpera, the ecosystem, structured into four interoperable layers, is already operational:

• AI Terminal (Super App): a ChatGPT-style dApp, deployed on the BNB Smart Chain (BNB Chain), with over 2 million registered users and hundreds of thousands of DAU (daily active users, metric not yet publicly disclosed) (data reported in 2025 and verified as of September 30, 2025) BNB Chain: over 2 million users.
• Agent Developer Platform: a marketplace where over 10,000 agents have been published, reusable in different compositions, as reported by industry outlets The Block.
• Model & GPU Platform: an infrastructure designed for model providers, data, and computing capacity, integrating Web2 and Web3 solutions ChainOpera.
• AI‑Native Blockchain Protocol: the Proof of Intelligence (PoI) protocol to measure, validate, and remunerate contributions, making the entire system verifiable and auditable. The technical specifications are described in the research sections of the whitepaper.

Proof of Intelligence: how performance and contributions are measured

The PoI is a verifiability protocol designed to attribute value to the efforts of agents and providers. It is based on three pillars:

• Measure: evaluation through standardized tasks (for example, accuracy, success rate, latency, cost per task, and resistance to adversarial inputs).
• Verification: use of on‑chain attestations, peer checks, and sample re‑executions to mitigate fraud and sybil attacks.
• Liquidation: distribution of rewards proportional to the value produced, with penalty mechanisms (slashing) in case of discrepancies.

In essence, an agent deposits the outcome of a task; independent verifiers check it with shared benchmarks and the blockchain records both the result and the related payout. This way, a quality economy is incentivized, not just the consumption of resources.

Impact: a cooperative and onboarding-oriented AI

The cooperative nature of the network allows agents and models to improve as data availability and computational capacity increase.

Each new contribution integrates into the system, enabling chains of specialized agents, as described in the whitepaper and public project communications ChainOpera.

The conversational interface facilitates access to DeFi and other on‑chain tools, making AI a natural bridge for non-expert users and generating a virtuous circle between usage, development, and interoperability.

Competition to Centralized Models: Where It Can Win and Where It Cannot

Decentralized systems focus on specialization and composition of agents, contrasting with large centralized models. Among the advantages are:

• Transparency and the possibility of auditing on executions and costs;
• Resilience against single points of failure and censorship;
• Incentives that reward high-quality contributions (data, models, or computational capacity).

That said, significant challenges remain related to the adoption of effective governance, the control of agent quality, and compliance with privacy and intellectual property regulations, as well as potential limitations due to latency and verification costs in real-time applications.

Context: how it positions itself compared to other projects

In the global landscape, Bittensor, SingularityNET, and Fetch.ai follow different paths: Bittensor incentivizes the provision of intelligence through network and model mining; SingularityNET offers a marketplace for AI services; Fetch.ai focuses on an economy of autonomous agents.

Yet, these differences highlight the positioning of ChainOpera, which stands out for an already active user terminal, an agent marketplace, and a native verification protocol (PoI), aiming to bridge the gap between development and adoption.

Roadmap: priorities and contribution opportunities

• Agent Social Network: development of multi-agent experiences and smoother interactions between humans and AI, to foster group collaborations.
• Verifiability: refinement of the PoI protocol through public benchmarks and even more robust attestations, to ensure transparency. A topic addressed in detail in AI Benchmark: complete guide to evaluating AI model performance.
• Developer platform: enhancing the scalability of agents, with applications in sectors such as market intelligence, automation, and on‑chain execution.

The roadmap highlights the areas where communities and partners can experiment and contribute to the validation of key components, maintaining an application‑led approach that aligns research and adoption in the long term.

In Summary

ChainOpera combines a blockchain-based AI platform, collaborative agents, and a Proof of Intelligence mechanism designed to measure and reward contributions.

With 2 million registered users, 10,000 agents published, and a product already live, the project offers a concrete path to decentralization of AI, while facing significant challenges in governance and quality control, as confirmed by recent industry analyses Cryptocurrencies and emerging projects: ChainOpera among the leaders.