Analyzed all transaction data from December x402, with nearly 60% used for practical work rather than speculation.

Author: E.H.

Translation: Deep Tide TechFlow

We analyzed all x402 transaction data from December 2025:

63 million payments

$7.5 million USDC

Over 1,100 projects involved

This is the first real-world case demonstrating that AI agents can provide large-scale payment services.

Below are our findings

A quick review of history:

HTTP 402 “Payment Required” status code was defined in 1997 but has never been practically used.

The reason is simple: credit card processing a $0.01 payment requires a $0.30 fee, resulting in a 3000% fee. Micro-payments were doomed from the start.

Until the emergence of stablecoins, which changed everything.

Overview of December 2025 data:

63 million transactions

$7.5 million USDC in circulation

64,000 independent buyers

10,000 independent sellers

Average transaction amount is only $0.12

In comparison: Visa’s minimum viable transaction is about $15, while x402 is handling payments nearly 100 times smaller and has achieved scaled operation.

The ecosystem is exploding:

Over 1,100 independent projects

Over 4,800 mainnet endpoints

6 active payment coordinators

Most service providers are small-scale, with no dominant players.

Translate this: Now is the critical moment to “grab territory.”

76% of service fees are $0.10 or less

Best range? $0.01-$0.10, accounting for 69% of all services.

This is the micro-payment domain—completely uncompetitive for credit cards.

AI agents pursue efficiency optimization; they pay on demand and are always billed per use.

So what are people building?

🥇 Data services: 31%

🥈 AI/Large Language Models (LLM): 25%

🥉 Blockchain: 15%

Not meme images, nor speculative projects.

But real infrastructure: APIs, analytics, reasoning services.

“Real money” has already entered the scene.

Base vs Solana—who is the winner?

Base: 53% (developer’s main platform)

Solana: 37% (preferred in production environments)

Base has Coinbase’s ecosystem; Solana wins with speed.

Which one to choose? It depends on your use case.

“Isn’t this all fake transaction volume?”

Let’s be honest:

Artemis data shows that 47% of transaction volume is for leaderboard boosting.

But—this only accounts for 14% of total transaction value.

Small transactions = cleaner fund flows.

The feasibility of micro-payments has been validated.

Unsolved billion-dollar problems:

What happens when AI agents make payment errors?

x402 has already solved the “how AI agents pay” problem.

But still unresolved:

Agent identity verification (KYA, Know Your Agent)

Reputation system

Dispute resolution mechanism

Who can solve these problems will be the ultimate winner.

63 million transactions per month

Average transaction amount: $0.12

The protocol is working.

The economic model is viable.

Developers are in place.

The only remaining question is: what will you build on this foundation?

Click to view the full report