Author: Robbie Petersen Source: X, @robbiepetersen_ Translation: Shan Oppa, Golden Finance

Understanding the Internet One framework for success is to view it through the lens of coordination. Fundamentally, we can attribute the success of the most valuable Internet applications to their ability to more finely coordinate human intent. Amazon reconciles commercial intent. Facebook, Instagram and Twitter orchestrate social intent. Uber and Doordash coordinate ride-hailing and delivery intentions. And Google coordinates information/search intent by matching queries to relevant web content.

It is becoming increasingly clear how artificial intelligence agents will underpin the next generation of logic evolution Coordination at scale. While today our “intentions” are carried out by searching, downloading and interacting with apps on the internet, it seems reasonable to assume that in the near future our “intentions” will be driven by an artificial intelligence that works together on our behalf proxy network to perform.

Importantly, this shift to an agent-mediated economy raises a fundamental question: What infrastructure will ultimately support this evolution?

In this article, we will:

(1) Discuss the use of artificial intelligence agents in encryption Bull and bear market cases of trading on track;

(2) Outline the logical path adopted;

(3) Explore where value will ultimately accumulate in this emerging agency economy.

Crypto Infrastructure

With the rise of the agency economy, blockchain is seen as a potential economic foundation for this new economic system, a view that has sparked widespread discussion . However, as with much of the emerging crypto space, the logic supporting this vision is often reduced to an overly singular narrative: “AI agents cannot have bank accounts, so will use crypto wallets.” This statement, while intuitive, ignores Understand the true value of cryptographic infrastructure.

In fact, AI agents are not completely incapable of using bank accounts. For example, via FBO ("For Benefit Of") account structure, which can open virtual sub-accounts for agents. This structure has been widely adopted by companies like PayPal to manage millions of sub-accounts. Stripe also recently announced support for proxy trading based on a similar structure. Therefore, from a technical and operational perspective, AI agents are perfectly capable of managing accounts through the traditional banking system.

Some argue that using traditional banking systems would undermine the autonomy of AI agents, but this argument is weak. Ultimately, the AI ​​agent's private key still needs to be managed by someone, and even if it is stored in a Trusted Execution Environment (TEE), it will be difficult to promote due to high cost and scalability limitations. Moreover, even if agents are fully autonomous, their core purpose is still to serve humans, so their autonomy is not absolutely necessary.

The real problem is that traditional payment systems have significant limitations in the following aspects, making it difficult for them to support the efficient operation of agent transactions:

1. Settlement time

•Traditional payment issues: Cross-border transactions usually take several days to complete settlement and are subject to batch Processing limitations make it impossible to meet the AI ​​agent's demand for real-time response.

•Blockchain solution: The public chain achieves almost instant settlement through atomic transactions without relying on counterparties, supporting 24/7 transactions without geography or restrictions on banking hours.

2. Global accessibility

•Traditional payment issues: The impact of the traditional banking system on the global There are significant barriers for developers, especially for the 70% of developers outside the United States, accessing payment systems is particularly difficult.

•Blockchain solutions: Public chains are naturally borderless and permissionless, allowing global agents to be deployed seamlessly without the need to establish traditional banking relationships. As long as you have access to the Internet, you can access the Internet.

3. Transaction economics

•Traditional payment issues: Fee structure of traditional payment systems (such as 3% plus fixed fee) makes small-amount high-frequency transactions uneconomical and limits the operational capabilities of AI agents.

•Blockchain solutions: High-performance public chains support low-cost, small-amount transactions, allowing agents to complete tasks with high frequency and low value.

4. Technical accessibility

• Traditional payment issues: Traditional payment infrastructure lacks programmatic APIs and is limited by strict PCI compliance requirements. For manual interaction, you need to rely on manual input or web forms, which hinders the realization of automated operations.

•Blockchain solutions: Blockchain infrastructure provides native programmatic access through standardized APIs and smart contracts, eliminating the need for screen scraping or manual input simulations, reducing compliance costs And improves automation reliability

5. Multi-agent scalability

•Traditional payment. Problem: Managing multiple AIs Segregating funds and accounts for agents is complex in traditional systems, adding to the burden of banking relationships and financial management

•Blockchain Solutions: Blockchain. Addresses can be easily generated, supporting efficient fund isolation and multi-agent architecture, and smart contracts provide flexible programmable fund management without the operational burden of traditional banks.

The value of blockchain is not only to solve the problem of "agents cannot have bank accounts", but also to provide agent transactions with settlement efficiency, global accessibility, economy, technical flexibility and scalability. These characteristics make the blockchain an indispensable infrastructure for the agent economy and provide strong support for the efficient operation and scale of AI agents.

The road to adoptionAlthough the technical advantages of cryptographic infrastructure are attractive, they are not necessarily necessary in the early stages of agent-driven commerce. Traditional payment networks, despite their many limitations, have become popular due to their large network effects and ability to facilitate global commerce. To drive adoption of new infrastructure, it must provide significant benefits beyond marginal improvements.

Looking ahead, AI Agent adoption is expected to progress in three phases, each embodying greater agent autonomy:

Phase 1: Human vs. Agency transactions (current stage)

We are going through this initial phase of emergence. A prime example is @perplexity_ai’s recently launched “Buy with Pro” feature, which shows how humans can complete transactions through AI agents. Their system allows AI Bots conduct product research for users, compare options, and execute purchases through a one-click checkout process with a traditional credit card or a digital wallet such as Apple Pay.

While in theory this process could use a crypto payment network, there is currently no significant reason why this is necessary. As @lukedelphi points out, at the heart of this issue lies the degree of autonomy required for the agent. In At this stage, agents have very limited autonomy. They do not manage resources independently, do not take risks, and do not pay for other services - they only serve as auxiliary tools to assist users in completing research and decision-making, and ultimately the user performs checkout operations. /p>

Until the agency economy reaches subsequent stages, the limitations of traditional payment networks will gradually become apparent, prompting increased demand for cryptographic infrastructure.

Chapter Phase 2: Transactions between agents and humans (rising)

In this phase, agents begin to independently initiate transactions with humans. This model is already used in some cases. Narrow fields are beginning to take shape, such as:

•Financial transactions: AI The trading system automatically executes the buying and selling of stocks or crypto assets.

• Smart Home: The system automatically purchases electricity based on time-of-use pricing to save costs. ="text-align: left;">•Supply chain management: automatic inventory replenishment based on demand forecasts

In the future, this kind of relationship between agents and humans. transactions will become more diverse and complex, such as:

•Payments and Banking: AI optimizes bill payments, cash flow management, detects fraud and disputes charges, automatically categorizes expenses and maximizes revenue with smart account management.

•Shopping and consumption: price monitoring and automatic purchasing, subscription optimization, automatic return claims, smart inventory management of household supplies.

•Travel and transportation : flightClass price monitoring and rebooking, intelligent parking management, shared travel optimization, and automated travel insurance claims settlement.

•Home management: intelligent temperature and power optimization, predictive maintenance scheduling, automatic replenishment of consumables based on usage patterns.

•Personal finance: tax optimization, insurance comparisons, portfolio rebalancing, bill negotiation with service providers.

Although these use cases are beginning to reveal the shortcomings of traditional payment networks, it is still theoretically possible to complete these transactions using proxy SDK structures such as Stripe. However, this phase will drive a shift from fixed periodic payments (e.g. monthly, annual) to usage-based pricing. Agents will optimize spend in real time, paying on-demand pricing for services such as API queries, computing resources, model inference, transaction fees, and more.

In this context, the shortcomings of traditional payment networks, especially the high fee structure (such as credit card payments), will become more prominent. The encrypted payment network shows great advantages in small and frequent transactions due to its high efficiency and economy.

The third phase: agent-to-agent transactions (future)

The final phase will trigger A paradigm shift in the flow of value in the digital economy. Agents will conduct transactions directly with each other, forming a complex autonomous business network. The prototypes of this stage have already begun to emerge in some speculative areas of the encryption market, but more complex scenarios will emerge in the future, such as:

•Resource market : Computing agents negotiate data storage locations with storage agents, energy agents trade grid capacity with consumption agents in real time, bandwidth agents auction network capacity to content distribution agents, and cloud resource agents perform real-time arbitrage between providers.

•Service optimization: the database agent negotiates query optimization services with the computing agent, the security agent purchases threat intelligence from the monitoring agent, the caching agent trades storage space with the content prediction agent, Load balancing agents work in conjunction with scaling agents.

•Content and data: content creation agent authorizes assets from media management agent, training data agent negotiates with model optimization agent, knowledge graph agent transaction verification information, analysis agent Purchase raw data from a data collection agency.

•Business operations: supply chain agency and logistics agencyCollaboration, inventory agents negotiate with purchasing agents, marketing agents purchase targeting data from audience agents, and customer service agents contract with dedicated support agents.

•Financial services: Risk assessment agents trade insurance policies with insurance agents, financial agents trade policies with investment agents, credit scoring agents sell verification to lending agents, liquidity agents Coordinate with market making agent.

This phase requires infrastructure support designed specifically for machine-to-machine (M2M) transactions. Traditional financial systems are built around human certification and supervision and cannot meet the needs of an agent-led economy. Stablecoins in the cryptocurrency system have become a key infrastructure for this economic form because of their programmability, borderlessness, instant settlement capabilities and support for micro-transactions.

Value Capture in the Agency Economy

The evolution toward an agency economy will inevitably produce winners and losers across the entire technology stack. In this new paradigm, several different layers become key points for value capture:

1. Interface layer: competing with the competition for end users in traditional payment environments Similarly, these actors may compete for the canonical interface layer over which end users express “agent intent.” These front-ends will gradually evolve from simple payment tools to comprehensive platforms that combine identity, authentication and transaction capabilities. There are several players that can capture value from this, including: (1) device manufacturers like Apple due to their hardware security and identity integration capabilities (2) consumer fintech super players like PayPal and Block’s Cash App Applications because of their large user base and existing closed-loop payment networks(3) like Chat-GPT, Claude, Gemini and Perplexity Such AI-native interfaces, since agent trading is a logical extension of their existing chatbots, and (4) existing crypto wallets, where they are able to leverage their crypto-nativeness as a first-mover advantage (albeit less likely).

2. Identity layer: A key challenge in the agency economy is distinguishing between human and machine participants. This is especially important in a world where agents begin to disproportionately manage valuable resources and make autonomous decisions. While Apple is in the best position here, @worldcoin is pioneering interesting solutions with its Orb hardware and World ID protocol. By providing verifiable proof of personality, Worldcoin could indirectly become one of the biggest winners from this structural trend, providing application developers with a platform that guarantees that all users are human. Although todayIt may be difficult to see why this is valuable, but it will become clearer in the future.

3. Settlement layer (blockchain): If blockchain can replace traditional railways as the normative settlement layer for artificial intelligence agents, then disproportionately promote agents The transformed chain will intuitively capture meaningful value at scale.

4. Stablecoin issuance layer: Given the liquidity network effect, it seems reasonable to assume that regardless of which stablecoin is used disproportionately by agents, Get meaningful value. USDC appears to be in the best position right now as Circle is rolling out developer-controlled wallets and stablecoin infrastructure to support proxy trading. However, it seems reasonable to assume that profits for stablecoin issuers will eventually compress as agents demand similar returns to businesses and humans.

In the end, the biggest losers may be those applications that cannot quickly adapt to the agent economy. In a world where agents, rather than humans, facilitate transactions, traditional moats will disappear. Humans make decisions based on subjective preferences, brand loyalty, and user experience, while agents optimize purely for performance and measurable results. This means that as the lines between applications and agents become increasingly blurred, value will flow to those companies that provide the most efficient, best-performing services, rather than those that build the best user interfaces or the strongest brands .

As competition shifts from subjective differentiation to objective performance metrics, end users (both humans and agents) will benefit the most.