Fetch AI Launches ASI:One and Business Tier for Agentic Web

1 hour ago7 min read

Fetch AI, the artificial intelligence infrastructure company founded by former DeepMind investor Humayun Sheikh, has launched a comprehensive trio of interconnected products designed to address the fundamental coordination problems plaguing today's nascent AI agent ecosystems. The release of ASI:One, Fetch Business, and the Agentverse directory represents a significant architectural bet on what the company terms the 'Agentic Web'—a future internet layer where autonomous AI agents from different organizations can securely interoperate to complete complex, multi-step tasks rather than merely providing recommendations.This vision stems directly from Sheikh's early experience at DeepMind, where he was among the first five people and the initial investor, providing him with a foundational perspective on how machine intelligence must evolve beyond pattern recognition toward executable action. The core innovation, ASI:One, functions as a personal AI orchestration platform specifically tuned for multi-agent coordination, storing user preferences in siloed knowledge graphs to handle context sharing and task routing across organizational boundaries.This contrasts sharply with conventional large language model applications that typically rely on retrieval-augmented generation or API calls to surface information but lack the persistent memory and coordination logic for transactional execution. Simultaneously, Fetch Business establishes a verification and discovery portal akin to ICANN domain registration for brand agents, allowing companies to claim official handles and verify identities through cryptographic challenges inserted into their existing website backends—effectively reusing the web's established trust layer.Completing this infrastructure stack is Agentverse, an open directory already hosting over two million agents, which serves as a universal discovery layer similar to DNS for the agentic ecosystem, enabling cross-platform interoperability regardless of the underlying development framework. This coordinated launch addresses what Sheikh identifies as the central limitation of current consumer AI: the inability to reliably execute actions requiring coordination across multiple business entities.The architectural philosophy here is distinctly modular, with Sheikh emphasizing that 'one large model isn't enough—you need specialists,' positioning ASI:One as the orchestration layer that coordinates these specialized agents. This approach reflects lessons learned from the early days of DeepMind, where Sheikh observed that agentic systems represented the most promising path toward practical artificial intelligence applications.The timing is strategically significant, arriving as major tech platforms struggle with siloed agent architectures and weak verification standards that have limited widespread adoption, with Sheikh noting that 'ninety percent of AI agents never get used because there's no discovery layer. ' Beyond mere coordination, the platform integrates payment pathways through partners like Visa and supported stablecoins, positioning digital transaction infrastructure as essential to the long-term vision of autonomous economic agents. This represents a substantial evolution beyond today's chat-based AI interfaces toward what could become a parallel web infrastructure specifically designed for AI-to-AI communication and commerce, potentially creating the foundational layer for an entirely new agent economy where intelligence becomes meaningful primarily through its capacity to act.

#Fetch AI

#ASI:One

#Agentverse

#AI Agents

#Enterprise AI

#Multi-Agent Systems

#featured

Stay Informed. Act Smarter.

Get weekly highlights, major headlines, and expert insights — then put your knowledge to work in our live prediction markets.