X402 has logged more than 169 million payments, yet the eight public MCP servers I operate have seen none that arrive without a test probe. The discrepancy raises a simple question: is the protocol truly being used, or are the numbers we hear built on a foundation that has yet to attract real spenders?

The headline numbers

Chainalysis—an analytics firm that tracks blockchain activity—reports that the x402 protocol has processed over 169 million payments. The same source says that industry heavyweights such as Stripe, Amazon Web Services, Google and Ripple have incorporated the protocol into their stacks. One striking metric: transactions exceeding a dollar’s value rose from 49 % to 95 % of all x402 activity within a single year. Those figures suggest that the protocol has moved beyond early-stage experimentation and is now handling genuine commerce.

What the logs on my servers tell me

I run eight publicly accessible MCP (multi-chain payment) servers. Two of them were recently upgraded to accept x402 payments as a low-risk experiment. Over a 30-day window I recorded every inbound request. The result was stark: zero payment attempts that originated from actual users or customer applications. All incoming traffic consisted of “scouts” – automated probes that map which endpoints can accept x402, verify handshake compatibility, and log the response code. None of these probes sent a payment payload; they merely asked, “Can I pay here?”

The logs also captured 22 distinct “payment offers” – requests that advertised the ability to receive x402 payouts. Those offers never turned into settled transactions. After the test period, the servers still show no organic payments – payments that arise without a preceding probe.

Why the gap matters

If the Chainalysis numbers reflect real commerce, the absence of organic traffic on publicly available nodes could be a temporary blind spot. If the numbers are inflated or derived from internal test environments, the protocol may be building an infrastructure that no one actually uses.

Two perspectives frame the debate:

  • Infrastructure-first view – New payment rails often appear before merchants or consumers adopt them. Developers build terminals, SDKs and API endpoints ahead of demand; early-stage probes may simply be preparation for a future wave of real transactions.
  • Bubble view – Developers and investors pour resources into a market that may never materialise. In that scenario, the probes are sunk costs – evidence of a speculative bet rather than genuine usage.

Both narratives are plausible, and the data I have collected does not yet tip the scales.

Signals that would settle the debate

I am watching for three concrete behaviors that would indicate real demand:

  1. Price inquiries – A probe that asks “what’s the fee?” instead of merely confirming capability suggests a buyer evaluating cost.
  2. Actual payment attempts after a 402 response – The HTTP 402 status signals “payment required.” If an agent receives that code and then retries with a payment payload, it shows intent to transact.
  3. Human-originated requests for per-unit pricing – A request that specifies a per-unit price rather than signing up for a subscription model points to a genuine commercial use case.

So far, none of these have appeared in my logs.

Takeaway

Chainalysis paints a picture of rapid growth for x402, but independent server logs still show a barren landscape of actual spend. The protocol could be on the cusp of a breakout, or it could be a speculative build-out waiting for demand that never arrives. The first genuine payment that bypasses a probe will end the debate. Until then, the thermometer stays at “watchful.”