AI agents—software systems that use AI to pursue goals and accomplish tasks on behalf of users—are proliferating. Think of them as digital assistants that can make decisions and take action to achieve the goals you set for yourself without needing step-by-step instructions: From GPT-powered calendar managers to trading bots, the number of use cases is growing rapidly. As their role expands throughout the economy, we must build the right infrastructure that will allow these agents to communicate, collaborate and trade with each other in an open marketplace.
Big tech players like Google and AWS are building the first marketplaces and trading protocols, but that begs the question: will they aim to extract massive rents through walled gardens once again? Agent capabilities are clearly increasing, almost daily, with the arrival of new models and architectures. What is at stake is whether these agents will be truly autonomous.
Autonomous agents are valuable because they offer a new user experience: the shift from software as a passive or reactive tool to an active, or even proactive, partner. Instead of waiting for instructions, they can anticipate needs, adapt to changing conditions, and coordinate with other systems in real time, without the constant participation or presence of the user. This decision-making autonomy makes them particularly suited to a world where speed and complexity exceed human decision-making.
Naturally, some are concerned about what greater decision-making autonomy means in terms of work and accountability – but I see an opportunity. When agents handle repetitive, time-consuming tasks and parallel what previously had to be done in sequence, they increase our productive capacity as humans, freeing people to engage in work that requires creativity, judgment, composition, and meaningful connection. It’s not imaginary, humanity has been there before: the arrival of business has allowed entrepreneurs to create entirely new products and previously unimaginable levels of wealth. AI agents have the potential to bring this capability to everyone.
On the intelligence side, truly autonomous decision-making requires an open-source and transparent AI agent infrastructure. The recent OSS release of OpenAI is a good step. Chinese laboratories, such as DeepSeek (DeepSeek), Moonshot AI (Kimi K2) and Alibaba (Qwen 3), have progressed even faster.
However, autonomy is not only about intelligence and decision-making. Without resources, an AI agent has few means to make changes in the real world. Thus, for agents to be truly autonomous, they must have access to resources and take care of their assets themselves. Programmable, permissionless, and composable blockchains provide the ideal substrate to enable agents to do this.
Imagine two scenarios. One where AI agents operate within a Web 2 platform like AWS or Google. They exist within the limited parameters set by these platforms, in what is essentially a closed, permissioned environment. Now imagine a decentralized market that spans many blockchain ecosystems. Developers can compose different sets of environments and settings. Therefore, the scope that AI agents have to operate is unlimited, globally accessible, and can evolve over time. One scenario sounds like a toy market idea, and the other is a real global economy.
In other words, to truly scale not only adoption of AI agents, but also agent-to-agent commerce, we need rails that only blockchains can provide.
The limits of centralized markets
AWS recently announced an agent-to-agent marketplace aimed at meeting the growing demand for off-the-shelf agents. But their approach inherits the same inefficiencies and limitations that have long plagued siled systems. Agents must wait for human verification, rely on closed APIs, and operate in environments where transparency is optional or non-existent.
To act autonomously and at scale, agents cannot be locked into closed ecosystems that restrict functionality, pose platform risks, impose opaque fees, or make it impossible to verify actions taken and their reasons.
Decentralization evolves agent systems
An open ecosystem allows agents to act on behalf of users, coordinate with other agents, and operate across services without authorized barriers.
Blockchains already offer the key tools needed. Smart contracts allow agents to perform tasks automatically, with rules embedded in the code, while stablecoins and tokens enable instantaneous and global value transfers without payment friction. Smart Accounts, which are programmable blockchain wallets like Safe, allow users to restrict agents in their activity and reach (via guards). For example, an agent may only be allowed to use whitelisted protocols. These tools allow AI agents to not only behave expansively, but also be contained within risk parameters defined by the end user. For example, this could include setting spending limits, requiring multiple signatures for approvals, or restricting agents to whitelisted protocols.
Blockchain also provides the transparency necessary for users to audit agents’ decisions, even when they are not directly involved. At the same time, this does not mean that all interactions between agents must take place on-chain. For example, AI agents can use off-chain APIs with defined access constraints and payments executed on-chain.
In short, decentralized infrastructure gives agents the tools to operate more freely and efficiently than closed systems allow.
This is already happening on Chain
As centralized actors continue to refine their agent strategies, blockchain is already enabling early forms of interaction between agents. Onchain agents already exhibit more advanced behaviors, such as purchasing predictions and data from other agents. And as more open frameworks emerge, developers are creating agents that can access services, make payments, and even subscribe to other agents, all without human intervention.
The protocols are already implementing the next step: monetization. With open marketplaces, individuals and businesses can hire agents, earn money from specialized agents, and create new services directly connected to this agent economy. Customization of payment models such as subscription, one-time payments or bundled plans will also be essential to meet different user needs. This will unlock a whole new model of economic participation.
Why this distinction is important
Without open systems, fragmentation shatters the promise of seamless AI support. An agent can easily complete its tasks if it stays within an individual ecosystem, such as coordinating between different Google applications. However, when third-party platforms are required (in social media, travel, finance, etc.), an open on-chain marketplace will allow agents to programmatically acquire the various services and goods they need to fulfill a user’s request.
Decentralized systems avoid these limitations. Users can own, modify, and deploy agents tailored to their needs without relying on vendor-controlled environments.
We have seen this work in DeFi before, with DeFi legos. Bots automate lending strategies, manage positions and rebalance portfolios, sometimes better than any human. Today, this same approach is applied as “agent legos” in industries such as logistics, gaming, customer support, and more.
The way forward
The agent economy is growing rapidly. What we build now will shape how it works and who it works for. If we rely solely on centralized systems, we risk creating another generation of AI tools that appear useful but ultimately serve the platform, not the person.
Blockchain changes that. It enables systems where agents act on your behalf, make money from your ideas, and connect to a wider, open market.
If we want agents that collaborate, transact, and scale without constraint, then the future of agent-to-agent marketplaces must live on-chain.
