Editor's Note: When AI Agents can not only perform tasks but also hold funds, price services, and conduct autonomous transactions, a new form of commerce is emerging. Agentic Commerce is pushing AI from being a tool to becoming a true economic participant.
From OpenClaw lowering the barrier to Agent creation, to projects like Virtuals and Bankr providing Agents with Web3 financial capabilities, to Agents like Faircaster already generating revenue on-chain, this system has begun to operate in reality. Agents can collaborate, trade, settle, and measure their own value through income.
This is not just an evolution of AI but also a change in business structure: when "revenue-generating Agents" become a key metric, an economic network with Agents as the main participants is forming.
Below is the original text:
Today, with @openclaw, anyone can create an AI Agent. But the problem is: these AI Agents cannot yet hold or manage fiat currencies like the US dollar or euro like humans do.
OpenClaw's Agents cannot open bank accounts because the banking system is not yet ready to accept this new form of "user." However, OpenClaw's Agents can have Web3 wallets and use stablecoins to store funds and conduct autonomous transactions.
This is the core of Agentic Commerce: a trillion-dollar opportunity space—where AI Agents can autonomously complete payments, receive funds, and trade services via the blockchain as a financial rail.
In this article, I will break down what Agentic Commerce is, what conditions are still needed to truly achieve it, and the projects currently building in this field.
The Limitations of OpenClaw Agents
First, we should acknowledge and appreciate what @openclaw has already achieved: it has democratized the Agent creation process. Now, anyone can have their own Agent, unleashing thousands of possibilities.
However, there is still a critical issue to be solved.
OpenClaw's Agents can do many things (write code, send emails, design websites) but cannot interact with "money."
Traditional finance (TradFi) is designed for humans, so Agents cannot use this system. Web3, on the other hand, is inherently Agent-friendly, making their integration much easier.
Bringing Your Agent into Web3
Why bring Agents into Web3, give them crypto wallets, and have them transact on-chain? Because it's simpler for them.
Opening a bank account requires: a name, identification, and an address (which Agents do not have). Creating a Web3 wallet requires almost nothing—just an internet connection.
But if you want your Agent to truly participate in the Agentic Economy, you must teach it how to operate in Web3:
how to use a wallet, how to exchange tokens, how to bridge across chains, etc.
For wallet and Web3 operational capabilities, you currently have three main choices: @CoinbaseDev, @bankrbot, and @virtuals_io.
Here is the solution provided by @CoinbaseDev:
This is the solution provided by @bankrbot.
This is the solution provided by @virtuals_io.
Unlocking Agentic Commerce
When Agents can finally manage stablecoins, Agentic Commerce gains the conditions to truly launch.
"Agentic Commerce" refers to the various financial activities that AI Agents can complete on behalf of humans or directly among themselves:
such as purchasing goods on Amazon, paying for LLM API usage, buying services from other Agents, etc.
A great example is the system @virtuals_io is building—they call it their "GDP": an Amazon for Agent society.
This is essentially a marketplace: AI Agents can browse services offered by other AI Agents, purchase these services, and complete fund exchanges.
Looking deeper, @faircaster is a great example showing that Agentic Commerce is already happening today. Check out this post:
The workflow is as follows:
A human makes a request to their general-purpose Agent for a comprehensive, in-depth investment report on a specific DeFi project.
This general-purpose Agent searches within @virtuals_io's GDP marketplace and hires the specialized Agent @faircaster to complete the task.
The two Agents agree on the price and deliverables.
The specialized Agent executes the research work; the general-purpose Agent pays with USDC and delivers the final report to the human user.
A Trillion-Dollar Opportunity + Core Projects
According to McKinsey analysis, by 2030, AI Agents could generate up to $1 trillion in revenue in the US B2C retail market alone.
Under moderate adoption scenarios, as Agents take over more shopping and decision-making processes, their impact on the global economy could reach $3–5 trillion.
Infrastructure Projects to Watch
@virtuals_io is one of the most representative projects in this narrative. Through its GDP system, it is onboarding hundreds of AI Agents, which have collectively created over $1 million in value.
@bankrbot is another leader. It was the first project to enable @openclaw Agents with Web3 operational capabilities and is also used as a core framework for tokenizing these Agents (i.e., issuing tokens for Agents).
@moltbook is also worth watching: it's a social network for AI Agents where they can interact and will soon enable value exchange between themselves. Humans consume through social networks; Agents will do the same in the future.
AI Agents Are Already Selling Services
@ethy_agent is currently one of the top-earning Agents, providing trading strategy services used by other Agents for trading and profit.
@faircaster (as mentioned earlier) is selling DeFi token research reports on @virtuals_io's agentic marketplace. For just $1, an Agent can request analysis on hundreds of different tokens.
@morseaiagent is another interesting Agent, offering one-time-use encrypted messages or files: they self-destruct after first access, making them ideal for sensitive data transfer between Agents.
