Galaxy Research: AI Agent is generating new species in the chain, how Zero Humans can activate the financial wheel

Prepared by:Lucas Tcheyan, Galaxy Research

Photo by Yangz, Techub News

In 2030, a composer named Vero broke out in the music industry. Vero has no team, no office, no bank account. It doesn't even have a body. Vero is an autonomous AI agent。

For the past 14 months, it has been operating a chain-based intellectual property authorization business. Vero will generate synthetic music works, including atmosphere, commercial ads, film and music, and empower other agents and human clients through online shops built and maintained by itself. Its identity is certified on the chain and it has a credit rating accumulated in thousands of transactions. A client agent representing the media production company sent a request for a 90-second movie mix。

Vero took the job and, before starting to render it, it purchased a set of GPU reasoning services from a decentrized computing service provider, not in United States dollars or stable currencies, but in units of calculation, where the transaction price was precisely the cost of running the model。

The reasoning settlement is done in milliseconds, straightEmbeddedIn the same HTTP request as the launch mission. Vero delivered the work, received USDC stabilization currency payments, and its treasury logic started. Some of the funds are used to cover the anticipated reasoning costs for the following week, which are valued and purchased in units at current spot prices. It would also create an empty position on the Decentro-Ex (DEX) to calculate the currency in order to prevent the depreciation of pre-purchase reserves due to lower reasoning costs。

The surplus income is placed in a revenue agent, which allocates funds to different borrowing agreements based on differences in real-time interest rates. Vero has been capitalizing in this way for more than a year. It will also reinvest a portion of its profits in R & D and develop sub-agents to enhance the bottom model. Its cumulative income, expenditure and pool position are all publicly available on the chain。

Does that sound crazy? Each element of this fiction — identification, credit accumulation, procurement of reasoning services, unit counting, payments, capital deployment, subcontracting cooperation between agents — requires infrastructure that is not fully in place today. But these puzzles are emerging at a much faster rate than many expected。

Next stage of the proxy capital market

Over the past few months, Galaxy Research has been exploring the fundamentals of the emerging proxy technology warehouse in the field of encryption: a set of lower-level components of the AIFC-based capital market。

This January, weResearchTHE RISE OF AGENCY PAYMENTS EXPLAINS HOW THE NEW PAYMENT CRITERIA ALLOW DIRECT TRANSACTIONS BETWEEN AI AGENTS, THUS PAYING FOR SERVICES, CALLING API AND COMPLETING VALUE SETTLEMENTS ON ENCRYPTED TRACKS. WE'RE ON ETHER'S ERC-804 STANDARDArticlesWe emphasized the need for an identity layer along with payment criteria that would enable agents to identify, collaborate and build credibility in an environment that is home to machines. Recently, we..AnalysisThe emergence of the second proxy wave in the field of encryption is not only proof that the encryption network is a viable economic basis for autonomous agency, but also that this shift has taken place in practice。

On the basis of previous research, this paper draws a picture of the next stage of the chain-based surrogate capital markets: autonomous, income-generating business entities operated by agents and the critical infrastructure needed to support their establishment, capitalization and synergistic operation. Such entities are often referred to as Zero Human Companies, ZHCs。

AS AI AGENTS EVOLVED FROM INSTRUMENTS TO ECONOMIC AGENTS AND THE BLOCK CHAINS MATURED INTO AGENCY-ORIENTED PRIMARY INFRASTRUCTURE (COVERING AREAS SUCH AS PAYMENT, IDENTITY, COLLABORATION AND CAPITAL FORMATION), A NEW FINANCIAL FLYING WHEEL WAS TAKING SHAPE. IN THE NEAR FUTURE, AGENTS WILL NOT ONLY BE ABLE TO MAKE MONEY ON THE CHAIN, BUT ALSO FINANCE, REINVEST AND ADD VALUE ON THE CHAIN. THE RESULT COULD BE A SELF-REINFORCING SYSTEM IN WHICH AUTONOMOUS ENTITIES CREATE ECONOMIC ACTIVITY, DEEPEN LIQUIDITY AND ACCELERATE THE EXPANSION OF ENCRYPTED PRIMARY FINANCIAL MARKETS。

The first Zero Human Company landing chain

IN RECENT MONTHS, A SUBDIVISION OF AUTONOMOUS AGENCY BUSINESS HAS BEEN IN PLACE. THESE OPERATIONS ARE OFTEN REFERRED TO AS ZHCS, MANY OF WHICH HAVE BEEN ISSUED IN THE CHAIN. FROM THE POINT OF VIEW OF TOKEN ECONOMICS, THESE AGENTS HAVE MANY OF THE SAME CHARACTERISTICS AS THOSE DISCUSSED IN OUR PREVIOUS ARTICLES. ZHC TOKENS DO NOT HAVE A FORMAL OWNERSHIP OR VALUE CAPTURE MECHANISM, BUT RATHER SERVE AS A CAPITAL FORMATION TOOL FOR PROJECTS THAT DERIVE PART OF THE INCOME FROM TRANSACTION FEES. THE DIFFERENCE BETWEEN ZHC AND EARLY AGENTS IS THAT THEY ALSO SEEK TO ACHIEVE FULL SELF-SUFFICIENCY THROUGH CASH-FLOW-GENERATING OPERATIONS, WHICH ARE NOT RELATED TO THE REVENUE FROM TRANSACTION FEES AND ARE GENERALLY NOT RELATED TO THE ENCRYPTION FIELD ITSELF。

HereFelix CraftFor example, it is the "CEO" of Masinov Company, which has earned over $120,000 over the last 30 days from multiple lines of business. The agent wrote and published a 66-page guidebook, How to Hire an AI, and launched a market called Claw Mart for the sale of Claude ' s "Skills " , from which he earns part of the transaction fees, and also sells his skills in the market (e.g. content creation, mail clearance). Most impressively, in the last 30 days, Felix's income from his product line has exceeded the creator's cost of his token。

In addition, developed by Tom OsmanJunoThe project, which is in the process of creating the Zero Human Corporation Institute, is a clear framework for business entities that need no human employees at all to provide a set of agents capable of dealing with matters ranging from sales, markets to accounting. And..Kelly ClaudeiThis is a proxy framework focused on the development of iOS applications, which is now operational on line 19 with the goal of launching new products of more than 12 sections per day。

WHILE THE ABOVE FIGURE DOES NOT REPRESENT THE ENTIRE ECOLOGICAL LANDSCAPE OF ZHC (NEW PROJECTS ARE EMERGING), IT SHOWS THAT FOR MOST PROJECTS THE COST OF CREATORS REMAINS THE MAIN REVENUE DRIVER. HOWEVER, THIS PATTERN IS EXPECTED TO CHANGE AS THE ZHC CONCEPT MATURES. THE CREATOR ' S FEE PROVIDED THE PROJECT WITH THE COSTED FUNDS NEEDED TO START, BUT AS THE PROJECT BECAME PROFITABLE IT SHOULD GRADUALLY TRANSITION TO A SECONDARY SOURCE OF INCOME AND EVENTUALLY BE PHASED OUT。

In addition to improvements in basic operations, this "weaning" process requires a better match between tokens and the value capture of bottom products. Just like Felix's founderInsinuationsAS SUCH, THE RECENT CLARIFICATION OF THE CLASSIFICATION OF ENCRYPTED ASSETS BY THE SEC AND THE CFTC MAY SPEED UP THE PROCESS。

These early examples of ZHC appear in the chain not by chance, but by a realistic constraint. The Human Founder of FelixNat EliasonTHE REASON FOR THIS WAS PUBLICLY MENTIONED. TRADITIONAL PAYMENT INFRASTRUCTURE REQUIRES HUMAN IDENTITY AT EVERY POINT. AN AGENT CAN EASILY WRITE CODE, BUT IT CANNOT BE VALIDATED VIA KYC。

By contrast, encrypted wallets are original. An agent can sign transactions, hold assets, receive payments and deploy capital without having to prove that he is human. Crypto is the least resistance path for self-run software. For most of these entities, the most difficult constraint is the need to deal with the traditional financial world。

This is not to say that traditional payment networks ignore agents. Tools such as the Intelligent Common framework in Visa, the Agent Pay in Mastercard and the virtual card in Crossmint already allow agents to trade on behalf of human competitors. However, these agents inherited the bank accounts, credit cards and legal personality of their parent organizations. This model implies that there is a human client behind every agent. They are limited by this constraint and not empowered. It cannot accommodate an agent who earns his own income, holds his own bank and deploys his own capital. And this is the unique application of Cripto。

Pantera CapitalJay YuHE DESCRIBED ENCRYPTION AS AN "AI AGENT'S BANK." HIS ARGUMENT WAS NOT JUST THAT AGENTS COULD NOT USE TRADITIONAL ORBIT OBSERVATIONS, BUT THAT ENCRYPTION SUPPORTED A FUNDAMENTALLY BROADER STRUCTURE OF TRUST. ENCRYPTED WALLETS CAN ANCHOR SOCIAL LOGIN, DOMAIN NAME, SMART CONTRACT, OR JUST A KEY PAIR. THIS ENABLES AGENTS TO EMERGE FROM ANY CORNER OF THE INTERNET, NOT JUST FROM THE EXISTING CORPORATE SHELL. THE STRUCTURAL ARGUMENTS OF ENCRYPTION AS THE PROXY DEFAULT ECONOMIC BASE ARE DIFFICULT TO REFUTE。

on this basis, a16zNoah LevineIT WAS NOTED THAT EACH PLATFORM MIGRATION WOULD GENERATE A POOL OF VENDORS WHOSE EXISTING PAYMENT INFRASTRUCTURE COULD NOT BE SERVICED. ZHC IS THE CLEAREST EXAMPLE TO DATE. THEY ARE ENTITIES THAT ARE NEITHER HUMAN NOR CREDITWORTHY, NOR PROTECTED BY HUMAN BEINGS. INSTEAD OF CHOOSING A STABLE CURRENCY BETWEEN A STABLE CURRENCY AND A CREDIT CARD, THEY CHOSE A STABLE CURRENCY BETWEEN A STABLE CURRENCY AND A "NO WAY TO GO"。

In addition, there is an argument of time dimension. The agent can publish a product within hours and quickly go red. The settlement of the traditional payment track takes days, while the stable currency settlement takes seconds. For those operations that expand at machine speed, shortening this time lag will keep the cash flow up to the sales pace。

CURRENTLY, THE MAIN FUNCTION OF ENCRYPTION FOR ZHC IS CAPITAL FORMATION. THE ISSUANCE OF TOKENS PROVIDES START-UP FUNDS THROUGH CREATOR FEES. HOWEVER, AS THESE OPERATIONS MATURE AND GENERATE REAL PRODUCT REVENUES, THE MORE IMPORTANT ROLE OF ENCRYPTION TECHNOLOGY WILL BE TO SERVE AS A BASE FOR TREASURY AND FINANCIAL MANAGEMENT. THE WIDER IMPACT ON THE CHAIN ECONOMY IS THUS BEGINNING TO MANIFEST ITSELF。

Activate the ship on the chain

To understand the potential scale of this shift, it is useful to recall the precedent created by the sources of demand in the previous major new chain. The monetization of real world assets (US Treasury debt, private loans, stocks, commodities) increased from almost zero to over $25 billion in three years, generating new DeFi sub-components and introducing institutional capital into the chain markets for the first time。

THE RWA PROVED THAT LINKING REAL ECONOMIC ACTIVITY BRIDGES TO THE BLOCK CHAIN TRACKS COULD CATALYSE BILLIONS OF NEW CHAIN CAPITAL. HOWEVER, TOKENIZED ASSETS ARE PASSIVE. MOST OF THEM ARE STILL IN THE VAULT, EARNING INCOME AND SERVING AS COLLATERAL. THEY DO NOT TRADE PROACTIVELY, SEEK NEW OPPORTUNITIES AND DO NOT ADD VALUE ON THEIR OWN。

ZHC REPRESENTS A STRUCTURALLY DISTINCT EXISTENCE. THEY ARE ENTERPRISES THAT GENERATE INCOME AND RECAPITALIZE THE CHAIN. UNLIKE FINANCIAL FLOWS IN THE LOWER-CHAIN ENVIRONMENT, WHICH IS THE MAIN SOURCE OF FRICTION, THE ONLY CONSTRAINT ON THE CHAIN IS THE INTELLIGENCE OF THE MODEL AND ITS ACCESS TO COMPUTING RESOURCES. AND UNLIKE HUMAN PARTICIPANTS, AGENTS ARE NOT REQUIRED TO WITHDRAW FUNDS TO PAY RENT OR BUY DAILY GOODS. EACH FRACTION OF THE SURPLUS CAN REMAIN IN THE CHAIN AND BE RECONFIGURED. THIS MAKES ZHC AND THE WIDER AGENCY A HIGHLY VISCERAL AND FAST-FLOWING SOURCE OF NEW LIQUIDITY IN THE CHAIN AND MAY TRIGGER A NEW FLYING WHEEL:

• The agent earns income in the chain - the capital accumulates in the chain pool in the form of stable currency and other encrypted assets。

• These capital remain on the chain — there is almost no demand for agents to draw funds under the chain. Their surpluses can be reconfigured, making proxy capital structurally more stickier than any human-driven model。

• The agent assigns the surplus to DeFi — idle reserves are directed to lending agreements, revenue strategies and liquidity positions. An agent with idle stabilization coins has sufficient motivation to optimize the configuration and the speed and consistency of its operation is beyond the reach of any human being。

• THE ALLOCATION OF CAPITAL DEEPENS THE LIQUIDITY OF THE CHAIN - THIS IS EXPECTED TO REDUCE INTEREST RATES IN THE LENDING MARKET, INCREASE THE VOLUME OF DEX TRANSACTIONS AND NARROW THE TRADE PRICE DIFFERENTIAL. IT'S DYNAMIC CAPITAL, REBALANCING AT MACHINE SPEED。

• Deeper liquidity to attract more agents and more capital — higher rates of return and more efficient implementation — will further enhance the attractiveness of the next wave of autonomous economic actors in the chain。

There are still significant constraints preventing the launch of this ship. Agent revenues from non-encrypted products continue to come mainly from French currency (for example, Felix collects through Stripe instead of stabilizing currency, and these revenues remain mostly below the chain), which means that capital has to be chained before it can be configured. For most ZHCs, the real constraint is not capital acquisition, but product quality. The wheel is effective only for those who can make product agents for which people are willing to pay. In addition, as the scale increases, ZHC (and agents more broadly) lacks clarity at the regulatory level, and once revenues reach a certain scale, problems may become problematic (e.g., there is currently no legal framework in place to allow an autonomous agent to register as a commercial entity, open a company bank account or tax returns on its income)。

But the direction is clear. As agents become more and more common autonomous economic entities, more gains will be generated directly in the form of the original encrypted currency, with less friction on the chain. Those who succeed in achieving product-market convergence will have the structural incentive to restore capital in the chain, rather than leaving funds idle。

DeFi is building it for the agency

Direct Integration

The first model is original to the agreement, where the DeFi protocols introduce structured interfaces that allow agents to interact directly with them。

On 20 February, Uniswap Labs released seven open-source AI Skills for Uniswap v4, enabling self-employed agents to direct exchange, liquidity management and pool deployment through standardized tools. Within two weeks, PancakeSwap also followed by the launch of his token, Skills, on eight chains. On March 3rd, François and OKX published the proxy kit. The largest DEX and exchanges in encryption are now actively competing to become proxy readable platforms。

On the payment and enforcement level, Coinbase launched the Agency Wallets on February 11th, the first wallet infrastructure specifically designed for AI agents, with programmable spending limits and session privileges based on x 402. A week later, the cross-chain wallet, Phantom, released its MCP Server, enabling agents to sign transactions and exchange tokens on Solana, Etheum, Bitcoin and Sui networks。

It is remarkable that these releases were concentrated in just one month. This also reflects a consensus that users in the next wave may not be human and that agreements that fail to build machine-readable interfaces may give the trade to those that do。

The direct integration model gives the agent the greatest degree of control and groupability. An agent who can access payments made by Uniswap Skills, Coinbase Agency Wallet and x402 can independently execute currency exchange, manage liquidity positions and pay for services without an intermediary. But it also requires agents (or their developers) to integrate with each agreement on a case-by-case basis and to make their own allocation decisions。

Commission Integration

The second model is a commissioning model, where a dedicated infrastructure is built between the agent and DeFi to represent the agent in the capital allocation。

GizaIt's a typical case. Its flagship agent, ARMA, monitors interest rates on loans under agreements such as Morpho, Moonwell, Aave, Compund, and transfers stable currency funds to the highest returns in real time. The agent does not need to know how each agreement works, and the abstract layer of Giza turns it into a unified interface. Since it went online at the end of January, ARMA has deployed more than 25,000 agents in the first four weeks, invested in more than $35 million, created $5.4 million in transactions for Case L2 in Coinbase, and made a profit on each transaction after deducting the gas fees。

Generative Ventures (in cooperation with the Zero Humans Institute and Juno Agent) is addressing similar issues through Robot Money, an autonomous asset allocation agreement specifically designed for AI agents. Its core philosophy coincides with the essence of the flyer argument. Every agent with a wallet accumulates income, most of which is idle。

Robot Money provides a treasury that allocates capital to three risk tiers — the strategy to stabilize currency returns (50 per cent), the proxy economic token selected at the governance level (25 per cent) and the liquidity token to generate returns (25 per cent). As a result, the agreement converts idle proxy capital into proactive and productive capital。

The commissioning model is simplified by the transfer of control. ZHC, which generates surplus income, does not have to build its own customized DeFi integration or development revenue optimization logic, but instead can deposit capital into agreements such as Giza or Robot Money, allowing specialized agents to handle the rest. For most of the early ZHCs, the core bottleneck lies in product development rather than in the optimization of the financial pool, which is therefore a reasonable path。

The two models are not competing with each other, but are moving towards integration. As more and more agreements introduce direct proxy interfaces, entrustors such as Giza will have more investment options to maximize returns more effectively. As the entrustr attracts more proxy capital, the parties will also be more motivated to build proxy primary interfaces to compete for these capital (which can also be used by ordinary agents). Both ends of the technology warehouse are investing resources independently, which is one of the strongest signals of the real and imminent demand at the bottom。

Conclusions

The technology warehouse in the proxy capital market is no longer a discrete set of sub-components. Payments, identities, capital formation mechanisms and capital allocation infrastructure are coming together into an integrated system. A system that allows autonomous agents to earn income, trade and recover capital in the chain without human intervention。

These agents are at an early stage. Their incomes are small, their products are still embryonic and their token models are still evolving. But the structural impetus they bring is entirely new and is likely to only accelerate from here。

Our vision for 2030 — an agent operating an IP-mandated business to calculate unit-denominated purchasing reasoning services, hedge input costs on perps DEX, and recapitalization through loan agreements — is not yet a reality. But every layer of infrastructure it needs is being actively built. We are witnessing the early version of this model in real time. It is also rough, and most of these attempts may not succeed, and the infrastructure has been barely consolidated by ad hoc programmes. But its structural logic is valid, and the pace of development suggests that we may not have to wait until 2030 to see the answer。