This week, everyone is helping AI open bank accounts.

AI hasn’t learned to spend money yet, but the people fixing banks for it are already here.

Author: David, Deep Tide TechFlow

On March 18, another blockchain mainnet launched.

Its name is Tempo, backed by Stripe and Paradigm. Stripe is one of the world’s largest online payment companies, handling $1.9 trillion in transactions last year; Paradigm is one of the biggest venture capital firms in crypto. Last year, they jointly invested $500 million in Tempo, valuing the project at:

$5 billion.

A $5 billion blockchain—no trading tokens, no DeFi, no memes. On launch day, Tempo’s most high-profile product was:

Machines paying machines.

That sounds a bit abstract; you can think of it as AI now needing to spend money at every step. Calling an API costs money, buying computing power costs money, pulling data from a database costs money…

But existing payment systems are all designed for humans. Bank accounts require ID, credit cards need facial recognition, Alipay needs mobile verification codes.

AI can’t pass any of these.

It can help you complete the entire workflow, but when it comes to paying, it has to stop and wait for a human to click “Confirm.”

So, along with the mainnet launch, an open protocol called MPP (Machine Payments Protocol) was also introduced, co-developed by Stripe.

In simple terms, it sets the rules for transactions between machines, including how to request payments, authorize, and settle.

The envisioned scenario is that AI can autonomously spend within preset limits, without needing human signatures for each transaction. On launch day, over 100 service providers integrated, including OpenAI, Anthropic, and Shopify.

But Tempo isn’t the only one doing this this week.

Within five days, Visa established a new department and released AI payment tools, Coinbase upgraded its payment protocol significantly, Mastercard spent $1.8 billion to acquire a stablecoin company, and Sam Altman’s World launched a toolkit for AI identity verification.

Five giants rushing through the same door in a week, eager to open bank accounts for AI.

Two paths, one door

Tempo is about settling payments for AI. But settlement is just one part of the payment system. For an AI agent to truly spend independently, it also needs payment tools, funding channels, and identity verification.

Here, traditional payment companies and crypto firms are competing in their own ways.

On the same day Tempo’s mainnet launched, payment giant Visa also moved. Its newly formed Crypto Labs department released its first product: Visa CLI, a tool that allows AI agents to initiate credit card payments directly from the terminal.

No API keys needed, no pre-registration required. When an AI needs to buy a service during a task, it can pay with a single command. Visa calls this “Command Line Commerce.”

Visa’s global card network connects hundreds of millions of cards and tens of millions of merchants. If AI payments can run on this existing network, there’s no need to wait for new infrastructure to mature.

Visa is extending its traditional path. Its competitor Mastercard chose a different approach: buying the road outright.

On March 17, Mastercard announced it would acquire London-based stablecoin infrastructure company BVNK for $1.8 billion. This is the largest stablecoin acquisition in crypto history.

The goal is straightforward: if AI payments are made with stablecoins, then stablecoins can flow through Mastercard’s pipeline.

On the crypto side, actions are equally intense.

Coinbase’s x402 protocol received a major upgrade, expanding payment scope from a few stablecoins to all ERC-20 tokens, and released the MCP toolkit, enabling developers to connect AI tools to payment networks with one click.

Though their starting points seem different, both are heading in the same direction: Traditional payment companies are embracing crypto, and crypto companies are embracing AI. Ultimately, crypto infrastructure is becoming the underlying pipeline for AI payments.

One more link remains. AI can spend money, but how do merchants know there’s a responsible human behind that AI?

On March 17, Sam Altman’s World released AgentKit, integrated with Coinbase’s x402. Its only function: to let AI prove that a verified human is behind each payment. Merchants can confirm someone is responsible for the transaction, but they can’t see who that person is.

In five days, five companies have secured each part of the process—settlement, channels, tools, protocols, and identity.

The AI cake is divided, leaving only the checkout counter.

Over the past three years, most of the AI industry chain has been claimed.

The model layer is dominated by OpenAI, Anthropic, Google, and numerous Chinese companies; compute power is tightly controlled by Nvidia; the application layer is a red ocean of programming assistants, search engines, and more…

Every layer is crowded, and barriers to entry are rising.

But the payments layer remains relatively vacant.

It’s not that no one thought of it; the timing just wasn’t right. AI agent payments require a prerequisite: AI must be capable of independently completing an entire task chain. If it only chats, without calling APIs, buying compute, or hiring other agents, then payments aren’t a necessity.

In the past year, this prerequisite has begun to materialize.

OpenClaw enables AI to directly operate computers; MCP protocol allows AI to connect to external services; major models’ agents are expected to make a significant leap in capabilities in late 2025. AI is shifting from a “dialogue tool” to a “work tool,” and working costs money…

The demand to spend is here, but the infrastructure to do so isn’t yet.

That’s why Stripe, Visa, Mastercard, and Coinbase are all stepping in. For traditional payment companies, this is their first chance to take the lead in the AI wave. They can’t build models or chips, but they’ve been handling payments for decades.

Visa’s global card network connects hundreds of millions of cards and tens of millions of merchants; Mastercard covers over 200 countries; Stripe processed $1.9 trillion last year. If every AI expenditure flows through these channels, the more capable AI becomes, the more they profit.

For crypto companies, the logic is different.

Coinbase CEO Brian Armstrong once said plainly: “AI can have crypto wallets, but can’t open bank accounts.”

Every step of the traditional financial system verifies “who you are”: bank accounts require ID, credit cards need facial recognition, each transaction needs SMS verification. AI is software, not a person—these hurdles it can’t pass.

But crypto wallets don’t need these. A private key is an account; for an AI agent, on-chain payments are the easiest route.

Whether crypto or not, AI payments will become a new foundational market. The only difference is which pipeline is more machine-friendly.

Roads are built, but cars haven’t arrived

At this point, it seems everything is ready—five giants are poised.

But there’s one number worth noting.

Coinbase’s x402 protocol is currently the earliest and most widely adopted AI payment protocol. According to x402scan, in the past 24 hours, total transaction volume was $65,400. 150,000 transactions, averaging less than five cents each.

What infrastructure supports this? Tempo’s $5 billion valuation, Mastercard’s $1.8 billion BVNK acquisition, Visa’s new department, Stripe’s protocol development.

Billions of dollars in infrastructure serving a market with daily transaction volume comparable to a street-side bubble tea shop.

This seems to be the norm for infrastructure businesses.

Before the internet bubble in 2000, telecom companies laid millions of kilometers of fiber optic cables. Afterward, they found only about 5% of global internet traffic used those fibers. Most of those companies went bankrupt, but the fiber still exists.

Ten years later, streaming video and mobile internet filled those pipelines. The builders didn’t profit, but the roads are real.

AI payments are now at this stage. The demand logic is sound: AI agents are becoming more capable, genuinely need to spend independently, and require a new financial infrastructure.

Everyone is at the starting line, but once the starting gun fires, it turns out only they are on the track.

When the first truly autonomous transaction by an AI agent will happen in your life might come sooner than expected—or it might be later.

The only certainty is that this battle has begun, and your and my wallets might be the last to know.