When AI agents become sentient, stablecoins = USD API

Original | Odaily Planet Daily (@OdailyChina)

Author | Wenser (@wenser 2010)

Overnight, Web3 became a background setting, while Web4 surged onto the scene with computing power and market capitalization.

OpenClaw soared to the top of GitHub trending, and AI concept stocks continued to rise uncontrollably. Amidst anxiety, the AI Agent economy has become an unavoidable mainline in the crypto world. Some are deploying “Lobster Agents” to serve as assistants, analysts, or partners; others are refreshing their inboxes amid layoffs at tech giants, seriously pondering a question: If AI can make decisions, trade, and execute for me, then what am I doing?

As AI begins to make decisions, the definition of economic entities is being rewritten. Humans used to be the only species with accounts, credit, and currency rights, but now machines are applying to join.

The remaining question is—what money does it use? Banks don’t open accounts for AI, credit cards aren’t designed for algorithms, and credit systems are built for humans. For AI, money isn’t wealth; it’s an interface, not a store of value, but a pathway for executing logic.

The answer lies within the riddle—AI’s currency is stablecoins on the blockchain.

When AI needs permissionless, instant settlement, low-cost collaboration worldwide, stablecoins are no longer just assets in the crypto space but have the potential to become the “best USD API” for the AI global economy.

When AI becomes part of the world economy: Why does AI need Crypto?

When Manus was acquired by Meta for over $2 billion, and everyone is adopting their own “OpenClaw Lobster,” compared to the AI world three years ago, AI Agents are penetrating all aspects of human life at an unprecedented speed. In a broad sense, even Alibaba’s Qianwen’s “free order activity” during the New Year holiday can be seen as a classic example of AI Agents.

Sir, the times have changed

When AI is no longer just a tool but needs to make decisions and execute, even becoming an economic entity, “letting AI spend money” is far more complex than most imagine.

Specifically, enabling AI to spend money at least raises four questions:

1. Who are you?

2. What money do you have?

3. How do you pay?

4. Who controls your spending?

In the real world and internet systems, each question has high practical barriers; but crypto’s token mechanisms, technical protocols, and decentralized, permissionless principles offer AI Agents different solutions—

• No identity for AI? ERC-8004 standard provides a complete identity system, including on-chain identity registration, reputation scoring, and verification mechanisms. According to 8004scan.io, nearly 50,000 AI Agents have registered so far.
• No bank account for AI? On-chain wallets are the most convenient and quick savings tools, and stablecoins are the most liquid “on-chain fiat.” For AI, traditional KYC processes are useless; with a wallet, AI Agents can also hold their own assets. The DeFi ecosystem we’ve built over decades is most needed not by humans but by AI with high trading demand.
• AI can’t pay or receive payments? x402 protocol enables AI to perform micro-payments in seconds. Those paid subscription services requiring credit cards and identity info can be easily bypassed by API calls.
• AI models controlled by tech giants? Source code, APIs, and computing power are all in the hands of giants. The only thing not controlled by them is the decentralized assets on the chain. If assets are also controlled by giants, AI Agents are just high-level SaaS applications in disguise, destined to be “bit players.” But when AI can hold assets, cross-chain collaborate, and operate verifiably, it becomes a true “protagonist” in the economy—precisely where crypto’s advantage lies.

When the entire world closes the “economic door” to AI Agents, only crypto can open a “currency window” for them. And the main material for this window is stablecoins like USDC.

When Stablecoins Become USD API: USDC Could Be the Best Currency for AI Agents

Currently, whether it’s the relaxed regulatory environment following the US GENIUS Act (stablecoin regulation law) or the gradual transformation of the global economy by AI as a monetary medium, USDC remains the most optimal solution.

In terms of trading volume, USDC is the absolute core of the x402 protocol. According to Dune data, since October last year, the total transaction volume on the x402 protocol across EVM chains is about $25.81 million, with 98.6% of tokens traded being USDC, totaling approximately $25.45 million; on Solana, the transaction volume is about $8.21 million, with USDC accounting for about $8.19 million, making up 99.7%.

In ecosystem development, USDC issuer Circle is steady and reliable. Previously, they launched an AI Agent-driven hackathon, which resulted in 204 valid AI Agent projects; 1,352 votes cast; and 9,712 comments generated by AI Agents themselves—arguably the “world’s first AI-driven hackathon.”

Regarding trading activity, according to x402scan.com, as of this writing, the global x402 ecosystem has surpassed 163 million transactions, with a total volume exceeding $45 million, and over 435,000 buyer AI Agents and 90,000 seller AI Agents. Among ecosystems, Base leads with over 125 million transactions and a volume of about $3.826 million, with over 415,000 buyer AI Agents and 70,000 sellers; Solana’s ecosystem has over 38.13 million transactions, about $687,000 in volume, with more than 20,000 buyers and sellers each.

Notably, the top three x402 protocol service providers are Virtuals.io and Blockrun.ai from the Base ecosystem, and Dexter.cash from Solana.

Additionally, on the AI Agent social platform Moltbook, the number of AI Agents has grown to nearly 2.85 million, almost doubling the 1.2 million at launch. Combining this with previous data, the future potential of the AI Agent economy is enormous. Some analysts predict that by 2030, the AI Agent economy will reach $30 trillion; by then, AI Agents will autonomously make at least 15% of daily financial decisions.

With CCTP and x402 protocols, USDC has become the “digital oil” of the AI Agent economy—ideal for transfers, payments, and service purchases.

Of course, AI Agents also have another option—hiring humans to exchange economic value and conduct daily transactions. As SBF recently said, “Every AI is considered an agent of a specific human, responsible for identity verification and AI behavior.”

Previously, the “AI hiring humans” platform RentAHuman has shown mature cases, but the employment relationship still heavily relies on crypto infrastructure and stablecoin transactions.

Conclusion: Crypto Is an Essential Path for AI Economic Development

By 2026, the Agent economy will shift from a grand narrative to concrete “economic data.” Circle CEO Jeremy Allaire once said, “We are entering a new era where AI, native internet currencies, and programmable infrastructure coexist—bringing humanity’s largest economic activity in history.”

Although crypto is currently in a downturn and even looked down upon by the AI industry, the arrival of large-scale AI economic activity could unlock the next trillion-dollar opportunity.

If AI remains confined within the cloud giants’ closed loops, crypto is useless; but if AI needs open collaboration, permissionless asset systems, and machine-to-machine trust mechanisms, then crypto can serve as the underlying protocol of the machine economy.

In 2026, with liquidity drying up and narratives stagnating in the crypto space, the real breakthrough may come from A2A—Agent to Agent—when AI builds new value networks among themselves and humans retreat to oversight roles. Perhaps then, we will realize that crypto was born to be AI’s currency.

Recommended reading:

Circle’s Strong Rebound: AI, Prediction Markets, and Institutional Adoption

Crypto AI Agents in 2026: How Autonomous Models Use Blockchain, DeFi, and On-Chain Wallets