Written by: a16z crypto

Translated by: AididiaoJP, Foresight News

AI Agents are rapidly evolving from auxiliary tools to genuine economic participants at a pace far surpassing other infrastructures.

Although Agents can now perform tasks and transactions, they still lack standardized methods to prove "who I am," "what I am authorized to do," and "how I will be compensated" across different environments. Identities cannot be migrated, payments are not yet inherently programmable, and collaboration remains isolated.

Blockchain is addressing these issues from an infrastructural level. Public ledgers provide verifiable evidence for each transaction that anyone can audit; wallets grant Agents portable identities; and stablecoins serve as an additional settlement layer. These are not future concepts; they are already available today and can help Agents operate as real economic entities in a permissionless way.

Providing Identity for Non-Humans

The current bottleneck of the Agent economy is no longer intelligence, but identity.

In the financial services industry alone, the number of non-human identities (automated trading systems, risk engines, fraud models) is already about 100 times that of human employees. As modern Agent frameworks (tool-calling large models, autonomous workflows, multi-Agent orchestration) are deployed at scale, this ratio will continue to rise across various industries.

However, these Agents are still essentially "unbanked." They can interact with the financial system, but cannot do so in a portable, verifiable, and inherently trusted manner. They lack standardized means to prove their authority, operate independently across platforms, or take responsibility for their actions.

What is missing is a universal identity layer—equivalent to an Agent version of SSL—that can standardize collaboration across platforms. Current solutions remain fragmented: on one side, there are vertically integrated, fiat-first stacks; on the other, crypto-native, open standards (like x402 and emerging Agent identity proposals); and there are developer framework extensions attempting to bridge application layer identities (such as MCP, Model Context Protocol).

Currently, there is still no widely adopted, interoperable way for one Agent to prove to another Agent: who they represent, what they are allowed to do, and how they will be compensated.

This is the core concept of KYA (Know Your Agent). Just as humans rely on credit records and KYC (Know Your Customer), Agents will need cryptographic signatures and attestations that bind them to subjects, authorizations, constraints, and reputations. Blockchain provides a neutral coordination layer: portable identities, programmable wallets, and verifiable proofs that can be parsed in chat applications, APIs, and marketplaces.

We have already seen early implementations emerge: on-chain Agent registries, wallet-native Agents using USDC, ERC standards for "minimal trust Agents," and developer toolkits that combine identity with embedded payments and fraud control.

But until a universal identity standard emerges, merchants will continue to block Agents at the firewalls.

Governing Systems Run by AI

As Agents begin to take over real systems, a new question arises: who truly has control? Imagine a community or company with an AI system coordinating critical resources (whether allocating capital or managing supply chains). Even if people can vote on policy changes, if the underlying AI layer is controlled by a single provider, which can push model updates, adjust constraints, or override decisions, then that authority is very fragile. The formal governance layer may be decentralized, but the operational layer remains centralized—who controls the model ultimately controls the outcomes.

When Agents assume governance roles, they introduce a new layer of dependency. Theoretically, this can make direct democracy more feasible: everyone can have an AI agent to help understand complex proposals, model trade-offs, and vote according to established preferences. But this vision can only be realized if Agents are genuinely accountable to the people they represent, are portable across providers, and are technically constrained to follow human directives. Otherwise, what you have is a system that superficially appears democratic, but is actually governed by opaque model behaviors that no one truly controls.

If the current reality is that Agents are primarily built on a few foundational models, we need a way to prove that an Agent is acting in the interest of users, rather than that of the model company. This will likely require providing cryptographic assurances at multiple levels: (1) the training data, fine-tuning, or reinforcement learning that the model instance is based on; (2) the exact prompts and instructions followed by the specific Agent; (3) its actual behavior records in the real world; (4) trustworthy guarantees that the provider cannot change its instructions or retrain it without the user's knowledge after deployment. Without these guarantees, Agent governance will degrade into governance by those who control model weights.

This is where cryptography can particularly play a role. If collective decisions are recorded on-chain and executed automatically, AI systems can be required to strictly adhere to verified outcomes. If Agents possess cryptographic identities and transparent execution logs, people will be able to check whether their agents are acting within limits. If the AI layer is user-owned and portable, rather than locked into a single platform, then no company can change the rules with a single model update.

Ultimately, governing AI systems is fundamentally an infrastructural challenge, rather than a policy challenge. True authority depends on building enforceable guarantees within the system itself.

Filling the Void of Traditional Payment Systems for AI Native Businesses

AI Agents are starting to purchase a variety of services—web scraping, browser sessions, image generation—stablecoins are becoming an alternative settlement layer for these transactions. Meanwhile, a new class of market for Agents is forming. For example, the MPP market from Stripe and Tempo aggregates over 60 services specifically designed for AI Agents. It processed over 34,000 transactions in its first week, with fees as low as $0.003, and stablecoins are one of the default payment methods.

What differentiates these services is how they are accessed: there are no checkout pages. Agents read schemas, send requests, pay, and receive outputs, all in one exchange. This represents a new class of identity-less merchants: just a server, a set of endpoints, and a price for each call. No front-end interface, no sales teams.

The payment rails to achieve this are already live. Coinbase's x402 and MPP take different approaches but both embed payments directly into HTTP requests. Visa is also expanding card payment rails in a similar direction, offering a CLI tool for developers to spend from the terminal, with merchants receiving stablecoins instantly on the backend.

Current data is still in early stages. After filtering out non-organic activities like botting, x402 handles about $1.6 million in agent-driven payments monthly, far below Bloomberg's recent reported figures of $24 million (citing x402.org data). But surrounding infrastructures are rapidly expanding: Stripe, Cloudflare, Vercel, and Google have already integrated x402 into their platforms.

Developer tools present a significant opportunity, as "vibe coding" expands the pool of people who can build software, the total addressable market for developer tools is also growing. Companies like Merit Systems are building products for this world, such as AgentCash—a CLI wallet and marketplace connecting MPP and x402. These products allow Agents to purchase the data, tools, and capabilities they need using stablecoins from a single balance. For example, a sales team’s Agent can call an endpoint while pulling data from Apollo, Google Maps, and Whitepages to enrich lead information, without the user leaving the command line.

This Agent-to-Agent commerce tends to utilize crypto payment rails (and emerging card-based solutions) for several reasons. First, there's the underwriting risk: traditional payment processors need to take on merchant risk when onboarding merchants, and a headless merchant with no website or legal entity is challenging to underwrite by traditional processors. Second, stablecoins have permissionless programmability on open networks: any developer can make an endpoint support payments without onboarding a payment processor or signing a merchant agreement.

We have seen this model before. Every transformation in business forms creates a new type of merchant that existing systems initially struggle to serve. Companies building this infrastructure are betting not on a monthly $1.6 million but on what that number will look like when Agents become the default buyers.

Repricing Trust in the Agent Economy

For the past 300,000 years, human cognition has been the bottleneck of progress. Today, AI is pushing the marginal cost of execution towards zero. As scarce resources become abundant, constraints shift. When intelligence becomes cheap, what becomes expensive? The answer is verification.

In the Agent economy, the real limitation to scale is our biologically constrained ability to audit and underwrite machine decisions. The throughput of Agents far exceeds human supervisory capacity. Due to high supervision costs and the lag in failures, the market tends to underinvest in oversight. "People in the loop" is rapidly becoming physically impossible.

However, deploying unverified Agents introduces compound risks. Systems ruthlessly optimize "agent" metrics while silently deviating from human intent, creating a hollow façade of productivity that obscures the accumulation of massive AI debt. To safely entrust the economy to machines, trust can no longer rely on human checks—trust must be hard-coded into the system architecture itself.

When anyone can generate content for free, verifiable sources become paramount—knowing where it comes from and whether you can trust it. Blockchain, on-chain proofs, and decentralized digital identity systems are changing the economic boundaries of what can be safely deployed. You no longer treat AI as a black box, but rather obtain clear, auditable historical records.

As more AI Agents begin to transact with one another, settlement rails and provenance proofs start to intertwine. Systems handling funds (like stablecoins and smart contracts) can also carry cryptographic attestations showing who did what, and who should be held responsible in case of issues.

The comparative advantage of humans will shift upward: from discovering minor errors to setting strategic directions and assuming responsibility when things go wrong. Lasting advantages belong to those who can cryptographically certify outputs, provide insurance for them, and absorb accountability in cases of failure.

Unverified scaling is a liability that accumulates over time.

Maintaining User Control

For decades, new layers of abstraction have defined how users interact with technology. Programming languages abstracted away machine code; command lines gave way to graphical user interfaces, followed by mobile applications and APIs. Each shift has hidden more underlying complexity but has always kept users firmly in the loop.

In the world of Agents, users specify outcomes rather than specific actions, while the system decides on its own how to achieve them. Agents not only abstract how tasks are executed but also who executes them. Users set initial parameters and step back, allowing the system to run autonomously. The user's role shifts from interaction to oversight; unless the user intervenes, the default state is "on."

As users delegate more tasks to Agents, new risks emerge: ambiguous inputs may lead to Agents acting on incorrect assumptions without users being aware; failures may not be reported, making clear diagnosis impossible; a single approval could trigger unforeseen multi-step workflows.

This is where cryptographic technology can be of assistance. Cryptographic technology has always aimed to minimize blind trust. As users delegate more decisions to software, Agent systems sharpen this issue and raise the rigor requirements in our designs—by setting clearer limitations, enhancing visibility, and enforcing stronger assurances about system capabilities.

A new generation of crypto-native tools is emerging. Scope delegation frameworks—such as MetaMask's Delegation Toolkit, Coinbase's AgentKit and Agent Wallet, and Merit Systems' AgentCash—allow users to define what Agents can and cannot do at the smart contract level. Intent-based architectures (like NEAR Intents, which have handled over $15 billion in accumulated DEX trading volume since Q4 2024) allow users to simply set expected outcomes (e.g., "bridge tokens and stake"), without needing to specify how to achieve them.