Author: Nancy, PANews
Recently, a dark horse has suddenly emerged on GitHub's star list. Hermes Agent, an independent AI agent framework that has only been open-sourced for a little over a month, has rapidly accumulated over 70,000 stars. In the history of the list, less than 0.0014% of projects have reached this number of stars.
Last weekend, Hermes Agent went viral online, with many developers abandoning "lobster" in favor of "Hermes." This wildly popular newcomer is backed by the Web3 forces.
In fact, Web3 is becoming an indispensable force for innovation and output in the AI ecosystem.
Behind the AI hit lies a group of Web3 veterans
As AI rapidly advances, a talent migration from the Web3 world is emerging.
Moltbook , an AI agent social network derived from OpenClaw, also generated a phenomenal buzz outside the crypto community. The product was eventually acquired by tech giant Meta. Its developer, Matt Schlicht, has a strong crypto background and is a serial entrepreneur.
The wildly popular Hermes Agent is backed by Nous Research, an open-source AI research organization. The project originated as an AI group within the Discord community. CEO Jeffrey Quesnelle previously served as the lead engineer for the Ethereum MEV infrastructure project Eden Network, giving Nous a Web3 DNA from the outset.
Alex Atallah, founder of OpenRouter, the world's largest AI model aggregation platform, was a co-founder of OpenSea. In 2022, when the NFT craze had not yet completely subsided, he chose to leave the Web3 field and turn to AI entrepreneurship.
Behind these rapidly gaining popularity are entrepreneurs with experience in the crypto industry. As the crypto market enters a downturn, Web3 talent will naturally flow to the AI field, which offers newer narratives and attracts more investment. Furthermore, the volatility of the crypto cycle has generally given crypto professionals greater resilience and iterative capabilities. Of course, during its previous boom, Web3 itself was a hub for global capital and talent, attracting a large number of outstanding developers. Now that AI has become a more explosive and long-term certainty-driven direction, some of these individuals will naturally shift their focus to AI.
However, at a more fundamental level, the rapid cold start and explosive community growth of these products are ultimately due to technological innovation and genuine market demand. Web3 experience is more of an added bonus than a core driving force.
Furthermore, Web3 talent is also demonstrating an advantage in AI security detection capabilities. In March of this year, AI giant Anthropic suffered a major security incident when approximately 512,000 lines of TypeScript code and over 1,900 files from its programming tool, Claude Code, were accidentally leaked. The first discoverer was Chaofan Shou, a researcher at blockchain security company Fuzzland and also an engineer at Solayer Labs. In addition, blockchain security organizations such as SlowMist also frequently monitor vulnerabilities and risks in AI systems.
Compared to traditional developers, Web3 security practitioners are more sensitive to attack surfaces, vulnerabilities, and anomalous behaviors. Furthermore, the industry's bounty hunter culture has cultivated highly skilled security researchers. This capability is gradually migrating to AI security auditing and risk discovery, injecting greater resilience into the AI ecosystem.
Web3 community, ideal seed users of AI
The Web3 community is naturally an ideal seed user for AI to become a productivity tool.
Take OpenClaw as an example. Before it truly broke out of its niche, the crypto community had already taken the lead, becoming one of the earliest users and dissemination nodes. Although market hype once bothered developer Peter Steinberger, the early popularity and participation brought to the project by crypto users cannot be ignored. Even offline lobster events in many parts of China were mostly initiated and organized by Web3 practitioners.
For example, to address the Sybil attack problem in the AI era, Worldcoin, co-founded by OpenAI CEO Sam Altman, combines biometrics, zero-knowledge proofs (ZK), and blockchain to verify real identities. However, without rewards, ordinary users may lack the incentive to participate. Building on this, Altman proposed the concept of Universal Basic Income (UBI) for the AI era, providing regular rewards to verifying users in the form of frictionless global tokens.
These cases show that, compared to ordinary people, Web3 users are often closer to tech optimists and early adopters. They are generally willing to accept higher learning costs and have a higher tolerance for the uncertainty of experimental products. Even before many products are mature, they are already willing to participate in testing, providing feedback, and even secondary dissemination. This characteristic perfectly matches the core user profile needed for agents to evolve into economic tools.
From a usage perspective, Web3 users are already accustomed to complex on-chain interaction processes, such as wallet signing, transaction execution, smart contract calls, and multi-protocol combinations. They have already overcome a significant learning curve, and therefore, when faced with agent hint engineering, token consumption, authorization mechanisms, or permission management, these users often possess stronger learning abilities and a higher acceptance rate.
At the same time, the crypto space inherently possesses financial incentives and speculative attributes, making information dissemination and user growth more likely to exhibit an exponential amplification effect, further accelerating the spread of the attention economy. Furthermore, the high activity of Web3 users on high-density social platforms such as X, Discord, and Telegram creates a highly efficient community environment for information dissemination, making it easier for new products or narratives to spread, develop, and amplify within a short period.
Although Web3 users are among the most concentrated experimental groups of early adopters of AI products, their long-term retention still depends on the product itself.
From ecosystems to infrastructure, Web3 is "invading" AI.
Web3's functionalities are also "invading" AI.
Nous's rapid expansion path clearly bears the hallmarks of Web3. Starting in 2024, Nous introduced the DePIN mechanism, incentivizing users to contribute idle GPUs and computing power for model training. This anticipated token airdrop directly stimulated community participation and accelerated the product's initial launch.
The scale and lineup of funding have further fueled community expectations for an airdrop. To date, Nous has raised over $70 million through its seed and Series A funding rounds, with the Series A round completed at a token valuation of up to $1 billion. The investor lineup is also heavily influenced by crypto, including Paradigm, Delphi Ventures, North Island, Solana co-founder Raj Gokal, and OpenSea founder Alex Atallah.
Today, the competition in AI has long surpassed simple technological contests. Nous's crypto-governance mindset and incentive design have expanded user engagement and ecosystem capabilities, laying the foundation for the wider adoption of Hermes Agent. It's important to note that Nous has not yet officially launched a token, and the airdrop remains uncertain.
A similar approach has emerged with OpenRouter, an AI ecosystem tool. As an aggregation platform that helps developers uniformly access and select AI models, founder Alex Atallah's initial entrepreneurial logic was that if creating a model is already so cheap, then there will be tens of thousands of models in the future, and they should each have their own internet presence. OpenRouter's role is similar to OpenSea's role as a liquidity aggregator in the NFT field.
This aggregation of liquidity in Web3 has allowed OpenRouter to directly benefit from the explosive growth of AI models. Currently, OpenRouter has integrated over 300 models, boasts over 5 million users globally, and its annualized revenue has increased from approximately $10 million last October to over $50 million. In terms of funding, OpenRouter has completed a $40 million funding round and is currently pursuing a new round of funding of approximately $120 million, bringing its post-money valuation to approximately $1.3 billion.
The cryptocurrency industry is also penetrating AI infrastructure. Crypto mining used to be a highly profitable business where miners could easily make a fortune by taking advantage of low electricity costs and high cryptocurrency prices. Even today's AI giant , Nvidia , made a fortune from GPU mining demand before officially entering the artificial intelligence field, tasting unexpected "windfall".
Despite the pressures of Bitcoin halvings, such as squeezed profits, rising energy costs, and price volatility, posing a survival challenge for many miners, AI is emerging as a new growth engine. Compared to the years-long process of building new data centers, crypto miners possess a unique advantage: their large-scale power contracts, data center infrastructure, cooling systems, land, and power grids are precisely the scarcest resources for AI data centers. This explains why leading mining companies like Bitdeer, Riot Platforms, and MARA Holdings are shifting their focus to AI/HPC (high-performance computing) .
In the fiercely competitive AI arena, Web3 is appearing more and more frequently, from upstream computing power and midstream framework models to downstream application scenarios.
