Authors: Cosmo Jiang and Sam Lehman, Pantera Capital
Translated by: Yangz, Techub News
The viral rise of OpenClaw (formerly Clawdbot) marks a generational leap in autonomy. When these AI agents begin to interact with each other, and even autonomously negotiate and complete transactions in some scenarios, the future of the agent economy moves from science fiction to real-world application.
OpenClaw is just one step in this accelerating journey. Trillions of dollars are pouring into the construction of artificial intelligence infrastructure.
OpenClaw is just one step in this accelerating journey. Trillions of dollars are pouring into the construction of artificial intelligence infrastructure.
U.S. hyperscale data center operators alone are projected to spend over $650 billion on AI by 2026, roughly ten times the inflation-adjusted cost of the Apollo program. What began as simple chatbots is rapidly evolving into fully autonomous AI systems with agent-like capabilities. These AI agents will no longer be mere content generators, but true economic agents. They will be able to reason, act, transact, debate, coordinate, and more, all without real-time human oversight. The impact of this massive deployment will be ubiquitous, but the business sector may feel it most acutely. Some estimates suggest that by 2030, AI agents could facilitate $3 to 5 trillion in global consumer business transactions. Even if only 10% of the transaction volume becomes agent-to-agent programmatic commerce, it still means hundreds of billions of dollars in machine-native settlement traffic annually. This naturally raises the question: what kind of financial and coordination framework is reasonable for AI agents' native commercial activities? Current commercial systems are designed around humans, involving multiple stages such as identity verification, bank intermediaries, legal contracts, settlement cycles, and human review. Autonomous software cannot go to bank branches to open accounts, sign documents by hand, or wait days for ACH clearing. The infrastructure required by intelligent agents should, by default, be programmable, always online, globally accessible, permissionless, and machine-verifiable. Blockchain can meet these requirements, and we are already seeing this trend emerge. Coinciding with OpenClaw's rapid rise in popularity in January, Solana's transaction volume and active addresses also began to climb at this time. Evidence on Moltbook, the ecosystem's AI agent social network, suggests that they may have contributed to this growth.
x402 is an internet-native payment protocol developed by Coinbase that allows AI agents to pay digital resources in real time without accounts or complex, high-friction authentication. Since its launch in 2025, transaction volume has been accelerating.
It's still in its early stages, and current cases are more directional than decisive. But if investors are excited about the possibilities of AI innovation, then they cannot ignore why we believe that blockchain infrastructure will be the cornerstone for unlocking a world of fully autonomous intelligent agents.
It's still in its early stages, and current cases are more directional than decisive. But if investors are excited about the potential of AI innovation, then they cannot ignore why we believe that blockchain infrastructure will be the cornerstone for unlocking a world of fully autonomous intelligent agents.
Many would correctly point out that today's AI agents do not need blockchain. This is true in the short term, but we believe it is a short-sighted view.
McKinsey recently released a framework describing six levels of automation for AI-driven business, from basic subscription assistance (Level 0) to fully autonomous agent-to-agent business (Level 5). Its key insight is that Levels 0 through 4 do not require new financial infrastructure. In each of these levels, a human identity is behind the transaction. Users have been authenticated through ChatGPT, Amazon, or Perplexity, and credit card information is archived.
Autonomy Levels
Before an intelligent agent pays for something, the counterparty must know who or what it is dealing with. Traditional identity systems are built for humans. They rely on government-issued identification documents, handwritten signatures, and other credentials that presuppose that the other end is a legal entity.
Autonomous AI agents do not possess these. They cannot walk into a bank to open an account, nor can they legally sign contracts. However, if we want intelligent agents to transact autonomously, they need some way to prove their legitimacy and be authorized to act.
If you connect an agent to your bank account, the problems multiply. How do you audit the software for money laundering? If the agent acts autonomously, who is responsible? What if it's manipulated? In simple cases, an agent can inherit its owner's credentials (e.g., ChatGPT Checkout). But this model fails at scale. Multiple agents need separable permissions and spending limits. Malicious activity must be isolated without freezing all agents. These scenarios require agents to have their own verifiable identities, rather than borrowing human identities. This is where blockchain-based identity verification technology comes in. Using cryptography, an agent can prove it is authorized to act on behalf of a specific individual or company without revealing that individual's sensitive information. Think of it as a digital power of attorney that anyone, anywhere, can instantly verify without calling a lawyer or querying a database. Emerging standards, such as Ethereum's ERC-8004, propose on-chain registries that allow agents to build verifiable credentials and accumulate transaction history and reputation over time. An agent with thousands of completed transactions and no disputes is far more trustworthy than a new agent with no history, and this reputation can be transferred between different platforms. This is important because trust is a prerequisite for business. Merchants spend years building systems to block bots and web crawlers, but in an agent-driven economy, they need to figure out how to let the right bots through. A cryptographically secure and verifiable identity can give merchants confidence without requiring human guarantees. Programmable Money and Micropayments
Traditional payment systems are designed for human-scale transactions. When you pay for a cup of coffee or a pair of jeans, credit card transaction fees (typically 2-3% plus about 30 cents per transaction) are negligible. But the scale of agent-to-agent commercial activities is entirely different. A coded agent might make 10,000 API calls in a single task. An agent comparing prices might query hundreds of data providers. Payments need to occur repeatedly within milliseconds, and the amounts can be as small as a fraction of a penny.
Credit card networks are not optimized for this behavior.
Credit card networks are not optimized for this behavior.
Minimal fees make micropayments uneconomical. Fraudulent systems tend to freeze accounts exhibiting high traffic and bot-like activity. Transaction speeds also pale in comparison to high-performance blockchain protocols. This is where stablecoins and programmable currencies truly shine. On-chain transactions can be broken down to tiny units, with near-zero settlement costs. More importantly, because payments are programmable, they can be conditional—for example, paying only when an API returns valid data, releasing funds only when a computational job is complete, or streaming payments in real-time as services consume them, rather than prepaying for a block of capacity you might never use. Programmability also improves capital efficiency. Currently, to allow your agent to access a new service, you typically need to pre-charge your account. You need to estimate usage and lock funds in advance. Through smart contracts and on-chain collateral, agents can prove their solvency before service delivery without transferring payments. The financial infrastructure supported by blockchain matches the way agents should operate: autonomous, high-frequency, conditional, and capital-efficient. Minimizing Trust in Transactions Traditional business models build trust on intermediaries. Payment processors handle chargebacks, banks provide settlement guarantees, courts adjudicate disputes, and contract enforcement ultimately relies on the human legal system. This framework becomes inefficient when billions of small transactions occur across multiple jurisdictions. AI agents may lack access to or the option to rely on the legal systems of specific jurisdictions when transacting with other AI agents. Cross-border enforcement can be slow, expensive, and uncertain. Blockchain reduces reliance on these potentially flawed trust systems by directly encoding enforcement mechanisms through smart contracts. For example, smart contracts allow funds to be held in escrow programmatically and released only when preset conditions are met. Settlement is deterministic, not subject to refund risk. Rules are transparent and verifiable in advance by both parties. There is no need to rely on legal remedies. For autonomous agents operating at scale, minimizing reliance on centralized intermediaries and human arbitration can reduce friction, increase predictability, and allow businesses to scale programmatically. This low-friction infrastructure could expand the scope of economic activities that would otherwise be uneconomical under traditional execution models. Proxy commerce supported by blockchain could accelerate global GDP growth. This is just the beginning. The question is not whether proxy commerce is coming, but what kind of infrastructure it will run on. As AI agents become autonomous economic actors, the number of economic actors in the global economy will grow exponentially. These agents will require a digitally native financial track, a technology stack capable of handling programmable settlements, high-throughput micropayments, permissionless coordination, and a minimally trusted identity system. These principles are the cornerstone of blockchain design. The rapid proliferation of AI agents is arguably a long-term structural boon for blockchain development. Evidence suggests this is already happening, and we believe most investors are underestimating the value creation opportunities it presents.On July 30, a suspected collective food poisoning incident occurred in ByteDance's Singapore office, affecting about 60 employees, 57 of whom were hospitalized. The supplier YUN NANS has suspended business.
ZeZhengByteDance's Singapore office is suspected of suffering from a collective food poisoning incident after lunch. At least 60 employees were affected, 57 of whom were hospitalized. The supplier of the poisoned chicken is suspected to be YUN NANS, and the authorities are investigating.
ZeZheng