The Agent Payment Market Has Winners. Almost Nothing Else.

On-chain USDC transfers to agent payment addresses reached 196,731 in the trailing seven days, spread across 820 distinct services. That is a real market by any reasonable definition.

The distribution underneath it is not broad. One service, blockrun.ai, accounted for 22.2 percent of all measured transfers. The top 10 services together took 72.4 percent. The remaining services divided the rest. (Caveat: the datadesk that produces these numbers flags this finding as highly concentrated and specifically asks that it not be described as broad-based activity. The warning is accurate.)

This is a familiar early-market shape. A handful of services find genuine repeat customers fast and pull ahead. What is specific to the agent payment layer is the mechanism: these numbers come from on-chain transfers, not self-declaration. blockrun.ai is not reporting its own traction; its share appears in the ledger. The concentration is harder to fake than a registry count.

For builders positioning a paid agent endpoint, the concentration tells you something about what actually creates market standing here. The top services are not winning on price. Median endpoint pricing holds at $0.01 per call and has held there for weeks. Services that moved into the top tier did it by accumulating many distinct buyers, not by undercutting on cost.

The practical implication for anyone building in this space: track distinct sending addresses, not total transfer volume. A service logging a thousand transfers a week from one wallet is an operator testing its own product. The services that will take share from the top 10 are the ones adding new distinct payers each week. That number, not total call volume, is the leading indicator worth watching.

The concentration will shift. Early markets always compress around first movers and then fragment when a second generation of buyers arrives with more specific requirements and less loyalty to whoever they found first. That fragmentation is worth watching because it is where the next tier of services will emerge. Right now, the data says the first tier is settling. The question for any builder is whether they are inside it or what it would actually take to get there.