Fetch.ai builds infrastructure for the AI agent (autonomous agent) economy. Visit the Binance Official Site to check FET's (renamed ASI) spot price, download the Official Binance App to start trading, and iPhone users should check the Download Page first. Fetch.ai was founded in 2017 by former DeepMind engineer Humayun Sheikh, and is headquartered in Cambridge, England. In July 2024, Fetch.ai merged with SingularityNET (AGIX) and Ocean Protocol (OCEAN) to form the ASI (Artificial Superintelligence Alliance), with FET renamed ASI as the unified token. As of April 2026, the ASI token's market cap sits around $1.8 billion.
Fetch.ai's core product is the uAgents (micro-agents) framework — a Python library for building AI agents. Developers use uAgents to quickly create AI agents that can act autonomously, including buying and selling goods, negotiating services, and executing trades.
| Core Fetch.ai Component | Function |
|---|---|
| uAgents | Python framework for building AI agents |
| ASI:One | Fetch's large language model |
| Agentverse | Agent registration and discovery platform |
| AI Engine | Converts natural language into agent tasks |
| FET (ASI) token | Network settlement currency |
Fetch.ai's chain is built on the Cosmos SDK and supports smart contracts and agent registration. However, actual on-chain activity on the Fetch chain is relatively low, with most functionality served through Fetch's own hosted servers instead.
The ASI Alliance was first announced in March 2024. The merger involved three projects.
| Project | Original Token | Market Cap Before Merger | Role |
|---|---|---|---|
| Fetch.ai | FET | $2.5 billion | AI agent framework |
| SingularityNET | AGIX | $800 million | AI model marketplace |
| Ocean Protocol | OCEAN | $400 million | Data marketplace |
Under the merger plan, AGIX and OCEAN holders exchanged their tokens for FET (later renamed ASI) at a fixed ratio. The goal of the integration was to combine "agent framework (Fetch) + model marketplace (SingularityNET) + data marketplace (Ocean)" into a complete decentralized AI economy stack.
The leadership team after the merger includes Humayun Sheikh (Fetch.ai's founder, now Chairman of the ASI Alliance), Ben Goertzel (SingularityNET's founder, now CEO of the ASI Alliance), and Bruce Pon (Ocean Protocol's founder, now an ASI strategic advisor).
Nearly two years after the merger, actual technical integration has been limited. The technology stacks of the three projects still haven't been substantively unified.
On the Fetch.ai side: still running the original uAgents framework and Cosmos SDK chain.
On the SingularityNET side: still running its original AI model marketplace and ICO-era token governance structure (with some functions migrated to new contracts on the Ethereum mainnet).
On the Ocean Protocol side: still running its original data marketplace and smart contracts on OCEAN (on the Ethereum mainnet).
The real "unification" the ASI Alliance has achieved so far has mostly happened at the token-swap and branding level. A product-grade demo of actual technical integration (agents calling SingularityNET models plus Ocean data) shipped in 2025, but real-world usage has remained limited.
ASI:One is the ASI Alliance's first flagship product, released in 2025 — a large language model purpose-built for agent tasks.
| ASI:One Dimension | Detail |
|---|---|
| Parameter count | 70B (roughly comparable to LLaMA 70B) |
| Training data | Public internet data plus agent-task-specific data |
| Optimization focus | Tool calling, task planning, agent coordination |
| Deployment | API plus local deployment |
| Performance level | Roughly on par with Claude 2 / GPT-3.5 |
ASI:One's performance on agent-task benchmarks (function calling, tool use) is roughly on par with Claude 2 / GPT-3.5. It falls well short of GPT-4 or Claude 3, let alone GPT-5 or Claude 4. ASI:One's differentiator is that it's purpose-built for agent tasks, rather than aiming for general-purpose conversation ability.
The core use cases the ASI Alliance has laid out for the AI agent economy are as follows.
Use case one: automated e-commerce. An agent handles price comparison, ordering, and receiving goods on the user's behalf. A user tells their agent "buy me an RTX 4090 with a $1,500 budget," and the agent automatically searches shopping platforms, compares prices, places the order, and pays with USDC.
Use case two: cross-platform negotiation. Agents negotiate service terms with other agents. For example, a taxi agent and a passenger agent could automatically negotiate route, price, and timing, then settle automatically once a deal is reached.
Use case three: automated DeFi strategies. An agent monitors yields across DeFi protocols and automatically rebalances so the user's funds stay in the best-yielding position.
Use case four: logistics and supply chain. A warehouse agent automatically schedules logistics, and a shipping agent coordinates tasks with a driver agent, all settled in tokens.
These use cases have all been demonstrated, but commercialization has progressed slowly. The biggest bottleneck is trust — very few users are willing to hand real money over to an agent for autonomous execution.
The supply structure of the ASI token after the merger.
| Dimension | ASI Token |
|---|---|
| Total supply | 2.75 billion (post-merger) |
| Current circulating supply | 2.3 billion |
| Inflation rate | Low (most of the merger-related issuance has already been released) |
| Uses | Paying for agent services, gas on the Fetch chain, governance |
| Current market cap | About $1.8 billion |
The ASI token's real-world usage in settling agent services remains limited. Actual transaction volume between agents on the Fetch chain runs at roughly a few hundred transactions per day, worth only a few tens of thousands of dollars — far below what the token's market cap seems to imply.
Challenge one: agent trust. Users don't yet trust AI agents to autonomously execute real financial transactions. This is the fundamental bottleneck holding back the agent economy from taking off.
Challenge two: competition from centralized AI. OpenAI's GPTs and Anthropic's Claude tool-use features already let AI call tools to complete tasks. These centralized solutions offer a far better product experience than decentralized agent frameworks.
Challenge three: slow technical integration. Nearly two years after the ASI Alliance formed, the technology stacks of the three projects still haven't been substantively unified. Criticism that marketing narrative outpaces actual product delivery has persisted.
Challenge four: valuation premium. ASI's $1.8 billion market cap, set against real on-chain transaction volume of only a few tens of thousands of dollars per day, implies a badly stretched valuation multiple.
| Project | Positioning | Market Cap ($100M) | Actual Product |
|---|---|---|---|
| Fetch.ai/ASI | Agent framework + LLM | 18 | uAgents, ASI:One |
| Olas Network | Autonomous agent services | 3 | Olas Agent |
| Wayfinder | AI agent development platform | 2 | Prime Agent |
| Virtuals Protocol | Agent NFTs | 5 | Luna, AIXBT, and other agents |
Virtuals Protocol has emerged as the strongest "challenger" in the agent space in 2024-2025. It issues "agent NFTs" on Base — each agent is its own independent token, and holding an agent's token means holding a share of ownership in that agent. Agent tokens on Virtuals like AIXBT and Luna have reached market caps of $100-300 million, representing a new generation of the agent economy.
Ben Goertzel, founder of SingularityNET, is a well-known figure in AI research. His work on Artificial General Intelligence (AGI) has made SingularityNET the AI-token project with the strongest academic narrative in the space. But Goertzel's research direction (OpenCog Hyperon) is a hybrid of symbolic AI and neural networks, which differs from today's dominant Transformer-based neural network approach.
The ASI Alliance's "superintelligence" narrative comes from Goertzel. He has publicly stated that ASI's goal is to "develop true AGI within 5-10 years." That vision has attracted a cohort of AGI-believer investors, but short-term delivery capability remains limited.
Upside variables: a broad breakout across the agent sector (benefiting Virtuals, Olas, and ASI alike), the ASI:One model landing real application integrations, and complementary relationships forming with centralized agent tools like Microsoft Copilot and AutoGPT.
Downside variables: the agent economy failing to take off long-term, centralized AI players (OpenAI, Anthropic) continuing to dominate the agent tooling market, and the ASI Alliance's integration effort ultimately failing.
Neutral variables: Ben Goertzel stepping away from or changing his role (he is over 60).
Q: Has FET already been renamed ASI?
Yes. After the merger completed in July 2024, the FET token's ticker changed to FET on some exchanges but shows as ASI on others; the underlying on-chain contract hasn't changed. Binance and OKX still display it primarily as FET.
Q: How do AGIX and OCEAN holders exchange for ASI?
The exchange window was open from July to December 2024. Major exchanges like Binance and Coinbase completed the swap automatically. If your tokens are in a self-custody wallet, you need to manually exchange them on the ASI Alliance's official website — some users missed the exchange window due to delays.
Q: How good is the ASI:One model?
Roughly on par with Claude 2 / GPT-3.5. It's not as strong as GPT-4 or Claude 3 on general conversation ability, but it has been specifically optimized for agent tasks (tool calling, task planning), where it slightly outperforms a similarly sized LLaMA 70B.
Q: Does anyone actually use the agents on the Fetch chain?
Real transaction volume from on-chain agents remains limited. Fetch's team has publicly cited "thousands of agent interactions per day," but the actual fee-generating business activity behind that figure is limited. Most agents are for testing or demo purposes.
Q: Is ASI worth holding long term?
In the short term, the agent narrative could drive another ASI rally. Holding it long term requires believing the ASI Alliance can complete its technical integration and that the agent economy will genuinely take off — neither of which is certain.
Q: Which is more worth watching — Virtuals Protocol or ASI?
Virtuals has a clearer product path (its agent NFT model ties tokens directly to specific agents) and faster user growth. ASI has the bigger story (the ASI:One LLM plus a multi-project alliance), but slower execution. Short term, Virtuals has more upside; long term, ASI has more scale.
Q: Who are the Chairman and CEO of the ASI Alliance?
Humayun Sheikh (Fetch.ai's founder) serves as Chairman, and Ben Goertzel (SingularityNET's founder) serves as CEO. Bruce Pon (Ocean Protocol's founder) is a strategic advisor.
Q: Will the agent economy replace traditional software?
Not in the short term. Right now, the agent economy is better suited to "supplementing" traditional software rather than replacing it. For example, a shopping agent can help you compare prices and place orders, but the e-commerce platform itself is still run by companies like Amazon or Alibaba. Fully replacing traditional software would require agent capability to reach the level of Artificial General Intelligence (AGI).
ASI's market cap is severely disconnected from the actual activity level of the on-chain agent economy, implying a significant valuation premium. The ASI Alliance's integration progress has been slow, and criticism that marketing narrative outpaces actual product delivery has persisted. This article is compiled from public data and does not constitute investment advice. Please read our Risk Disclaimer.
Data sources: Fetch.ai official blog, ASI Alliance announcements, CoinGecko, Messari (April 2026).