The data shows that the infrastructure race for agentic commerce is already producing winners, with Anthropic's Model Context Protocol running on over 10,000 public servers and pulling 97 million monthly SDK downloads.
Looking at on-chain metrics, it's clear that the current protocols, such as Agent-to-Agent and Universal Commerce Protocol, are solving the problem of how agents connect, coordinate, and initiate payments, but they do not address the issue of who decides if the work was actually done.
The Missing Primitive: Escrow
Statistically speaking, the current protocols are robust answers to the question of whether a purchase was sanctioned, but they say nothing about whether the purchased outcome materialized.
The payments incumbents building around agentic commerce are solving authorization, not verification, with protocols like AP2 and x402 enabling automatic payments over HTTP, but not determining whether the paid-for outcome materialized.
- The current protocols are solving the problem of how agents connect, coordinate, and initiate payments.
- None of the current protocols address the issue of who decides if the work was actually done.
- ERC-8183 proposes a programmable escrow primitive that holds funds conditionally until an evaluator attests that the deliverable has cleared.
The Power Center Nobody Named
The evaluator role is where the proposal becomes politically interesting, with the potential for a trust bottleneck and the need for durable answers to the question of who controls evaluation.
The data shows that the adoption numbers suggest the surrounding layers are moving faster than verification, with Gartner predicting that 33% of enterprise software applications will include agentic AI by 2028.
- Gartner predicts that 33% of enterprise software applications will include agentic AI by 2028.
- Deloitte pegs the global agentic AI market at $8.5 billion in 2026, rising toward $35 billion by 2030.
- IBM and NRF reported that 45% of consumers already use AI during buying journeys, including 41% for product research.
Our Take
As a data-driven analyst, I believe that the key to success in agentic commerce lies in solving the verification problem, and ERC-8183's proposed programmable escrow primitive is a step in the right direction.
However, the question of who controls the evaluation and the potential for a trust bottleneck remains, and it will be interesting to see how this plays out in the coming years.
