Written by: Thejaswini M A

Translated by: Bitpush

Visa's empire is built on "human participation," and humans are withdrawing.

Visa's entire business model is a gamble on human behavior. It concerns human consumption habits and psychology.

The points you accumulate, the fraud protection you rely on, the coveted Centurion black card, the “zero liability policy” that gives you peace of mind when using a card at foreign ATMs—these exist not because "moving money" is inherently difficult. Rather, it’s because humans are anxious, status-driven, and not good at reading terms and conditions. Visa has built a $500 billion company on this "cognitive gap."

AI agents, on the other hand, lack these traits altogether.

They do not accumulate points. They do not feel safer because of fraud protection. They do not aspire to hold black cards. They have one directive: complete the task. And when the task involves payment, the agent will do the calculations that humans are eternally too lazy to work out—the cheapest path, the fastest settlement, the lowest fees. Every time, automatically, devoid of any emotion.

Last month, a Substack article titled “The Global Intelligence Crisis of 2028” caused Visa's stock to plummet 4%, Mastercard to drop 6%, and American Express to sink 12%.

The report was interpreted as a "scenario analysis," not a "prediction" (as stated in the original text). But the market does not care. The technical assertion itself is not important; the problem is that by 2027, agents will bypass existing clearinghouses and use stablecoins for settlement. Visa spent 50 years crafting a perfectly precise product for a customer base that is being replaced.

In the “machine-to-machine” (M2M) business world, a 2% to 3% interchange fee rate is an unmistakable target. Citrini Research's line encapsulates the core argument. It’s not that AI will destroy Visa tomorrow; rather, the fee structure on which Visa built its empire has always been a tax on irrational human behavior, while agents are absolutely rational. That is the essence of it all.

What exactly is Visa selling?

To understand why this is crucial, you must grasp what the interchange fee actually funds.

When you buy something with a credit card, the merchant pays the card networks and issuing banks a fee of 2% to 3%. This money funds your rewards points, fraud protection, shopping insurance, and dispute resolution. The entire consumer value proposition of credit cards is funded by merchants, who pass on the cost to everyone by slightly raising prices. It is a beautifully stable system that has operated for 50 years because the “human” in the transaction is willing to pay for it all—just not directly.

AI agents do not need these. They do not require dispute transactions and do not want cash back. Those safeguards that prove high fees reasonable are fundamentally protections against human error, human fraud, and human impulse. Once the human is removed from the transaction, the logic of these fees completely collapses.

American Express (Amex) is the most typical version of this issue. Its customers are high-income, high-spending, aspirational elite cardholders. Its rates are higher than Visa or Mastercard precisely because its customers are willing to pay for status and privileges. This entire model assumes a human is consciously making decisions, choosing Amex over Visa for access to airport lounges. But agents will not choose Amex. Agents will only look for the cheapest option that can accomplish the task. In a world where cards are held by software, the so-called “high-end tiers” simply do not exist.

Agent-driven businesses avoiding fees pose a significant risk to banks and single-issue card issuers reliant on card business. These institutions are heavily reliant on the 2% to 3% fee revenue stream and have built entire business units around merchant-subsidized rewards programs. Visa and Mastercard have networks that can transition, while issuers that have built profit and loss models around fees and rewards will have nowhere to go.

A week of "everyone shipping at once"

The Citrini report and the rollout of various infrastructures just happened to collide in the same three-week window.

Tempo officially launched its mainnet on Wednesday. Stripe and Paradigm's payment blockchain (designed for high-volume stablecoin settlements) launched alongside the “Machine Payments Protocol.” This is an open standard that allows AI agents to autonomously process payments without humans needing to sign off at every step. The protocol introduces the concept of "Sessions": humans authorize a spending limit once, and then agents make continuous micropayments based on consumption data, computing power, or API calls. This is "money's version of OAuth." Humans authorize the budget, and agents are responsible for spending, with no need for a card at every step.

Anthropic, DoorDash, Mastercard, Nubank, OpenAI, Ramp, Revolut, Shopify, Standard Chartered, and Visa are all listed as design partners for Tempo. The entire payment and business stack is acknowledging this structural change.

On the same day Tempo launched, Visa's cryptocurrency division introduced a command line interface (CLI) tool for AI agents to make payments directly from the terminal, without an API key, account, or human authorization for each transaction. Visa defines it as “Command Line Commerce”—machines transact without human intervention.

Cuy Sheffield (@cuysheffield)

“I’m excited to share Visa CLI, the first experimental product from Visa Crypto Labs. Check it out and request access here: visacli.sh”

—March 18, 2026

Mastercard agreed to acquire stablecoin infrastructure startup BVNK for $1.8 billion. Circle launched Nanopayments on the testnet: sub-cent, gas-free USDC transactions designed for AI agents to pay for usage-based APIs, with no account or credentials required. Sam Altman's World project (formerly Worldcoin) launched AgentKit, allowing agents to carry cryptographic credentials proving they represent real humans, directly integrating them into Coinbase's payment path, enabling the platform to verify agents' identities without obstructing legitimate business activities.

In my view, what occurred this week is that companies are racing to become the “new Visa,” beating Visa to the punch before it realizes what it has lost.

The obvious paradox

Now, one point that hasn't been clearly articulated is that Visa is not sitting idly by.

It is participating in Tempo's machine payments protocol and has established Visa Crypto Labs, with its head of crypto explaining in Fortune how agents will use card pathways to make payments through new standards. Mastercard has invested $1.8 billion in stablecoin infrastructure. Stripe has acquired Bridge and Privy. Existing giants understand this shift and are trying to position themselves before the new infrastructure fully arrives.

Visa's argument is: it can expand its pathways to capture these agent transactions before establishing new pathways that render Visa irrelevant.

This argument is not obviously wrong. Stripe processed $1.9 trillion in total transactions in 2025, up 34% year over year. These companies are not shrinking. The channel advantage of card networks is not easily replicable. I admit I am a bit hesitant to say this out loud because historically, whenever someone makes this argument, a new product is released that makes the speaker look foolish.

So, the flaw in this argument is that Visa's channel advantage is built on merchant relationships and consumer trust. Merchants accept Visa because consumers carry Visa; consumers carry Visa because merchants accept it. The entire flywheel operates on the premise that there is a “human” in the transaction. Once agents become primary buyers in significant business categories, the flywheel will slow down. Agents have no brand loyalty and no wallets. What they have are budgets and directives. The cheapest, fastest path will always win their business, and the switching cost is zero.

I want to accurately describe our status because narratives often outpace data.

Despite the ecosystem around the x402 protocol being valued at approximately $7 billion, on-chain data shows that the protocol processed only about $28,000 per day last week, most of which came from testing rather than actual commercial transactions. This number is completely not on the same scale as the amounts Visa processes daily.

@artemisanalytics

The transaction volume for x402 has surpassed 50 million transactions. Each transaction amount is extremely small, but the transaction count indicates that the infrastructure is being used and developers are building on it. The merchant side (service providers accepting agent payments) is growing. This is what payment networks look like when they are in their infancy.

McKinsey predicts that by 2030, AI agents could mediate $3 trillion to $5 trillion in global consumer commerce. This valuation could be correct or overly optimistic. Undoubtedly, agent-driven commercial activity has not yet scaled. Merchants building agent-native services, companies deploying agents as primary buyers, and volumes that can truly pressure the existing fee economy are all still under construction.

The reason Citrini's report frightened the market is that it simulated a series of credible events. Mastercard's first-quarter earnings report for 2027 will not attribute the slowdown in transaction amounts to “agent-driven price optimization.” It’s not time yet.

This shock will first occur in the realm of micropayments for AI infrastructures, not standard consumer commerce.

An agent completing a research task could call specialized data APIs hundreds of times in a single session. The cost of each call is just a few cents. Over a week, these calls might generate $40 in revenue for the service developer. Card networks cannot process such transactions—the minimum transaction cost (economic model) does not work, the merchant onboarding process does not work, and the fee structure does not work. This type of business could never operate on Visa's pathways from the start. It requires something brand new, and x402, Nanopayments, and Tempo are building it.

Disruption of consumer commerce, according to Citrini's model, even if it does happen, is a future event. It requires agents to be able to take on a significant portion of discretionary spending, which in turn requires consumers to confidently hand over purchasing decisions they currently make to agents.

Visa is being disrupted by a higher-quality customer—the customer has no need for everything Visa once prided itself on. That 2-3% interchange fee is not a transaction tax; it is a tax on the irrationality of human nature. Agents are absolutely rational.

How do I know this will work? Because Visa spent $1.8 billion this week to ensure it does not get left behind.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。