Your AI just made several payments while you read that headline — you approved none of them, Visa processed none of them, and for some in crypto circles that’s the point: machine-to-machine micropayments could become the backbone of a new internet economy. Why this matters - Coinbase founder Brian Armstrong recently predicted there will soon be more AI agents than humans making transactions online. Binance founder Changpeng Zhao went further, saying agents will eventually make a million times more payments than people — and that most of those payments will be in crypto. - The argument isn’t just hype. It’s structural: banks require identity verification and KYC for accounts, which software agents can’t provide. A crypto wallet, by contrast, only needs a private key — no identity hurdles, no compliance delays. That makes wallets a far easier on-ramp for autonomous agents. But the economics are the sharper edge AI agents don’t shop like humans. When an agent executes a task — researching, coordinating logistics, or compiling a report — it often fires off dozens of tiny API calls in a single session. Each call might pay a fraction of a cent for compute, a real-time data feed, web scraping, or a specialist sub-agent. Those micropayments don’t fit the business model of card networks. Concrete example: imagine an agent writing this article - It queries a real-time news API to verify a tweet ($0.002), fetches on-chain volume data ($0.004), checks press releases ($0.001), consults a finance model for Visa protocol details ($0.003), then pays a tool to draft the text. Six microtransactions total under two cents on protocols like x402. - Stripe’s minimum per-transaction fee (~$0.30) would make running those six payments through a card network more than 100x the value of the payments themselves — economically absurd. Enter x402 and other crypto-native solutions x402 is Coinbase’s open payment protocol that embeds stablecoin payments into HTTP requests so an agent can hit a paywall, pay in USDC, and keep working — no human needed. Backers include Cloudflare, Circle, AWS, and Stripe. Google’s open agent payments standard lists x402 as a settlement layer. The promise: seamless, tiny, immediate payments for machine-to-machine interactions. Where this could be used - Healthcare: an agent managing claims pays per document retrieved. - Logistics: procurement agents auction and pay for freight slots in real time. - Media: crawling agents pay per article indexed instead of negotiating broad licenses. - Finance: trading agents pay a model fractions of a cent per risk signal consumed. Reality check: infrastructure is ahead of demand Adoption is still nascent. CoinDesk reports x402 processes roughly $28,000 in daily volume, and analytics firm Artemis says about half of observed transactions are artificial activity rather than genuine commerce. The kinds of merchants x402 targets remain relatively rare. Traditional finance is responding Card networks and banks aren’t sitting idle. Visa launched a Trusted Agent Protocol in October, and Mastercard completed Europe’s first live AI-agent bank payment through Santander’s regulated infrastructure recently — both running on existing card rails with cryptographic verification added. That shows regulated rails can adapt to agent payments too. The likely outcome We’re probably headed toward a split: regulated consumer and enterprise commerce stays on card rails, while high-frequency, ultra-low-value machine-to-machine payments — agents hiring agents, paying per API call, buying compute — shift to stablecoins because the economics demand it. Which of the two worlds will be larger over time remains the open question. For now, the race is on between legacy rails retrofitted for agents and crypto-native micropayment rails built from the ground up. Read more AI-generated news on: undefined/news

#AI #CZ #USDC