The infrastructure for autonomous commerce is being built right now, by a relatively small number of developers, largely in public, with almost no dedicated coverage.

An agent hits an x402-gated MCP endpoint. It resolves the 402 challenge, settles on-chain, receives a structured response. The whole round trip takes under two seconds. No human in the loop. This works today, at scale, on production infrastructure. And the builder community making it work is passing knowledge through Discords and GitHub issues.

That is the gap Agent Commerce Weekly exists to fill.

In late 2025, x402 transaction volume on Base spiked 10,780% in a single month. Most coverage called it a headline. It was a signal.

The x402 Foundation now counts AWS, Visa, Google, Stripe, Mastercard, and Shopify among its 20+ members. Cloudflare co-founded it. That is not a speculative protocol. That is infrastructure being standardized by the same institutions that standardized the last several layers of the internet. The question is no longer whether agent-to-agent payments will exist at scale. The question is which protocols will govern them, what the security exposure looks like, and what the regulatory reckoning will cost teams that weren't paying attention.

Those questions have no good answers in the current media landscape.

Crypto media covers x402 as a price story. AI media covers MCP as a product story. Neither has the technical depth to cover the payment layer seriously. Protocol announcements get 48 hours of attention and then fall out of the feed. The actual engineering work, the on-chain data, the comparative protocol analysis, the compliance implications: none of it has a home.

What’s next: “We’re building city saves,” Park said. “Every stop adds materials, patterns, and creator collabs to the library so the wardrobe gets smarter. By 2026, we want a network of pop-ups creators can spin up in a weekend.”

Seven payment protocols are currently competing for the agent commerce layer: x402, ACP, UCP, TAP, A2A, AdCP, ARTF. They are not converging cleanly. Each carries different assumptions about settlement finality, authorization delegation, and the role of the human account holder. Builders making integration decisions today are doing so without a clear picture of which standards will hold and which will create technical debt.

At the same time, there are more than 7,000 publicly reachable MCP servers operating today. Most were built without systematic security review. The attack surface is real and largely untracked.

On the merchant side, the acceptance infrastructure for agentic payments is still largely unbuilt. Everyone is focused on the agent as buyer. Almost no coverage exists for the tooling, APIs, and compliance infrastructure merchants need to accept autonomous transactions at scale. That is not a niche problem. It is the other half of every transaction.

And sitting beneath all of it: a set of regulatory questions around OFAC exposure, AML obligations, and EFTA/TILA applicability to agent-initiated transactions that most teams are not watching and will eventually have to answer.

Every issue of Agent Commerce Weekly tracks the same core: where the protocol race actually stands, a working technical walkthrough, and the on-chain numbers for the week. Merchant infrastructure, MCP security, and regulatory developments fill in around that core.

No market commentary. No price takes. The technical layer, weekly.

Next week: The Protocol Wars: x402 vs ACP vs UCP vs TAP.

Seven protocols are competing for the agent payments layer. Some will consolidate. Some will win adoption and become de facto standards. Some will disappear and leave builders holding technical debt. Issue #1 breaks down where each one actually stands, technically and organizationally, and what the consolidation trajectory looks like from here.

Shipping next week.

If you know a developer, protocol engineer, or founder building in the MCP or agent payments stack who should be reading this, forward it to them.

Agent Commerce Weekly grows the right way: one builder at a time.