The "bank card" of AI has caught the attention of the giants

Original author: David, Deep Tide TechFlow

On March 18, another blockchain mainnet went live.

It's called Tempo, backed by Stripe and Paradigm. Stripe is one of the largest online payment companies in the world, handling $1.9 trillion in transactions last year; Paradigm is one of the largest venture capital firms in the crypto industry. The two companies jointly invested $500 million in Tempo last year, valuing the project at:

$5 billion.

A $5 billion blockchain that doesn't speculate on coins, doesn't do DeFi, and doesn't issue memes. On the day of the mainnet launch, Tempo's most prominently announced product is:

Letting machines pay machines.

This sounds a bit abstract; you can understand it as AI needing to spend money at every step. Calling an API costs money, purchasing computing power costs money, pulling a batch of data from a database costs money...

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

AI can't pass any of these.

It can help you complete an entire workflow, but when it comes to the payment step, it has to stop and wait for a human to press "confirm."

Thus, along with the mainnet launch, an open protocol called MPP (Machine Payments Protocol) was launched, co-authored by Stripe.

In simple terms, it sets a set of rules for transactions between machines, including how to request payment, how to authorize, how to settle, etc.

The envisioned scenario is that AI can autonomously spend money within a preset limit without needing a human signature for every transaction. On the launch day, over 100 service providers were already integrated, including OpenAI, Anthropic, and Shopify.

But Tempo wasn't the only one doing this this week.

Within five days, Visa established a new department to release AI payment tools, Coinbase's payment protocol underwent a major upgrade, Mastercard spent $1.8 billion to acquire a stablecoin company, and Sam Altman's World released a toolkit specifically for AI identity verification.

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

Two paths, the same door

Tempo is focused on helping AI settle payments. But settlement is just one part of the payment system. For an AI agent to truly spend money autonomously, it still needs payment tools, funding channels, and identity verification.

In this regard, traditional payment companies and crypto companies are both vying for the cake in their own ways.

On March 18, the same day Tempo's mainnet went live, payment giant Visa also took action. The newly established Crypto Labs department released its first product: Visa CLI, a tool that allows AI agents to initiate credit card payments directly from terminals.

No API keys are needed, no prior registration is required; if AI needs to buy a service while running a task, it can pay with a single command. Visa calls this "command line commerce."

Visa's global card network connects billions of cards and tens of millions of merchants. If AI payments can run on this existing network, it doesn't need to wait for any new infrastructure to mature.

Visa is extending the old path. Its competitor Mastercard chose a different approach: buying the road directly.

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 the history of the crypto industry.

The purpose of this acquisition is straightforward: if AI payments are to be made with stablecoins, then stablecoins should flow through my pipeline.

On the crypto-native side, the actions are equally intense.

Coinbase's x402 protocol completed a major upgrade, expanding its payment range from several stablecoins to all ERC-20 tokens, while also releasing the MCP toolkit, allowing developers to easily integrate AI tools into the payment network.

While the starting points seem different, what they are doing points 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.

There is still one step left. AI can spend money, but how do merchants know if there is a responsible person behind the AI spending?

On March 17, Sam Altman's co-founded World released AgentKit, which integrates with Coinbase's x402. Its sole purpose is to allow AI to prove that a verified real person stands behind the payment. Merchants can confirm that someone is responsible for the transaction, but they cannot see who that person is.

In five days, five companies have positioned themselves in every step: settlement, channels, tools, protocols, identity.

The AI cake is divided, only the cash register remains

In the past three years, the positions in the AI industry chain have mostly been taken.

The model layer is dominated by OpenAI, Anthropic, Google, and a host of Chinese companies; computing power is tightly controlled by Nvidia; the application layer is a red ocean from programming assistants to search engines...

Every layer is crowded, and the competitive barriers are getting higher.

But the payment layer is still relatively vacant.

It's not that no one has thought of it; it's that the timing hasn't arrived. AI agent payments have a prerequisite: AI must first have the ability to independently complete an entire task chain. If it can only chat, doesn't need to call APIs, doesn't need to buy computing power, and doesn't need to hire other agents to work, then payment is not a necessity.

In the past year, this prerequisite has slowly started to be established.

OpenClaw allows AI to directly operate computers, the MCP protocol enables AI to access external services, and the capabilities of major models' agents are expected to break through in the second half of 2025. AI is transforming from a "conversation tool" into a "work tool," and working requires spending money...

The demand for spending has arisen, but the infrastructure for spending does not yet exist.

This is why Stripe, Visa, Mastercard, and Coinbase are all taking action at the same time. For traditional payment companies, this is their first opportunity to gain a home-field advantage in the entire AI wave. They can't build models or create chips, but payment is something they have been doing for decades.

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

For crypto companies, the logic is somewhat different.

Coinbase CEO Brian Armstrong previously stated directly: "AI can have a crypto wallet, but it can't open a bank account."

Every step of the traditional financial system confirms "who you are." Opening a bank account requires ID, getting a credit card requires facial recognition, and every transaction requires a text verification code. AI is software, not a person; it cannot pass any of these checkpoints.

But crypto wallets do not require these. A private key is an account, and for AI agents, on-chain payments are the path of least resistance.

Whether crypto or not, AI payments will be a new infrastructure-level market. The difference lies in whose pipeline is more suitable for machines.

The road is built, but the cars haven't come

At this point in the story, it seems everything is ready, and the five giants are in position.

But there is one number worth a glance.

Coinbase's x402 protocol is currently the earliest implemented and most widely adopted AI payment protocol. According to data from x402scan, in the past 24 hours, the total transaction volume of the entire ecosystem was $65,400. With 150,000 transactions, the average transaction was less than 50 cents.

What infrastructure supports this number? Tempo is valued at $5 billion, Mastercard spent $1.8 billion to acquire BVNK, Visa specifically established a new department, and Stripe personally drafted the protocol.

Infrastructure valued in the billions serves a market with daily transaction volumes comparable to a street-side milk tea shop.

All infrastructure businesses seem to be in this norm.

On the eve of the 2000 internet bubble, telecom companies laid millions of kilometers of fiber optic cables under the sea. After laying them, they found that global internet traffic only utilized 5% of it. Most of those companies went bankrupt, but the fiber optics remained.

Ten years later, video streaming and mobile internet filled those pipelines. The builders didn't make money, but the roads were real.

AI payments are now at this stage. The logic of demand is valid: AI agents are indeed becoming increasingly capable, needing to spend money autonomously, and requiring a new financial infrastructure.

Everyone is at the starting line, but after the starting gun fires, they find that the track temporarily only has themselves.

As for whose path will ultimately succeed, when the first true autonomous transaction of an AI agent occurs in your life may happen faster than everyone expects, or it may happen slower than everyone expects.

The only certainty is that this battle has already begun, and you and I may be the last to know.