Some weeks you cover a story. This week, the story covered us.

We published 14 original analyses in five days. LinkedIn impressions hit 84,658, up 118 percent from last week. And across every article, one thread kept pulling tighter: the gap between announcing agentic payments and actually wiring money through them is closing faster than anyone expected.

Here is everything that mattered.

The biggest infrastructure story of the week was Stripe unveiling its Machine Payments Protocol, a settlement layer purpose-built for autonomous agents. MPP is Stripe's answer to Coinbase's x402, and the two protocols represent fundamentally different visions: Stripe wants agents on card rails, Coinbase wants them on crypto rails.

We followed that thread into Coinbase itself, examining why the company is betting its future on agent payments. The x402 protocol, Base chain integration, and USDC settlement infrastructure form a coherent thesis. Whether that thesis survives contact with enterprise procurement is another question entirely.

Then the SEC provided unexpected clarity. Our analysis of the new stablecoin guidance showed regulators explicitly carving out space for stablecoins as settlement instruments. For agentic commerce, that is a green light that did not exist two weeks ago.

Nvidia's GTC conference dominated the middle of the week. We covered the security gap in Nvidia's agent infrastructure: the company is building the compute layer for agentic AI but leaving security and payments architecture to others. That is either a smart division of labour or a dangerous blind spot.

The broader picture landed in our deep dive on the $450 billion hardware pipeline now reshaping global payments. When data centre construction costs are settled through the same banking infrastructure that processes your coffee purchase, payments infrastructure becomes AI infrastructure. The two are no longer separable.

Google and Anthropic are taking opposite approaches to owning the agent layer, and we mapped both strategies in one analysis. Google is building the platform. Anthropic is building the model. Both believe they will win the same market. Only one of them is right.

Walmart's Instant Checkout flop was our most-read piece of the week, pulling 27,500 LinkedIn impressions alone. The article examined what happens when the world's largest retailer ships an agentic commerce feature that does not work at scale. The answer: it damages credibility for everyone building in the space.

That credibility theme carried into our Trust Gap analysis, which asked a harder question: does agentic commerce have a credibility problem that technology alone cannot fix? When Walmart, a company with functionally unlimited engineering resources, ships a broken checkout agent, it changes the conversation for every startup pitching the same capability.

Our analysis of the credit union agentic dispute gap examined why the credit union sector faces the same dispute infrastructure challenges as banks but with fewer resources. The piece was updated this week with additional regulatory context: the GAO confirmed that no federal banking regulator has issued standalone guidance on agent-initiated transactions, and the U.S. Treasury's AI Risk Management Framework is the closest any federal body has come to addressing autonomous agents in financial services.

The Andreessen Philosophical Zombie Investment Thesis explored a related angle: what happens when the biggest AI investor in the world needs humans to be shallow for his trillion-dollar thesis to work?

Three pieces this week focused on companies actually building the connective tissue.

Santander and Visa went live with agentic payments in production, not a pilot, not a sandbox. Live transactions with real money. We broke down what their architecture reveals about how traditional banking infrastructure adapts to autonomous agents.

Plaid wired open banking to AI, creating a bridge between regulated financial data and agent-driven commerce. The implications for identity, consent, and data access in agentic workflows are significant.

Our Agentic Commerce Demo to Live piece tracked the full journey from proof of concept to production deployment, identifying the five specific infrastructure problems that kill most agentic commerce projects before they reach a live customer.

This was also a building week for Major Matters itself. Our AI Tools Directory grew from 25 to 56 reviewed tools, more than doubling in a single week. Byline Pro, our content QA pipeline, received significant updates to error handling and link extraction. And we crossed 84,000 LinkedIn impressions for the week, a 118 percent increase that tells us the intersection of payments, AI, and commerce is finding its audience.

We have the Visa tokenization and invisible checkout analysis ready to publish. Expect continued coverage of the settlement layer war as Stripe MPP and Coinbase x402 start competing for the same enterprise customers. And we are watching regulatory developments closely, including new EU guidance on AI agent liability in financial services and the evolving US regulatory landscape for agentic payments.

Which of these stories changed how you think about your own payments or AI strategy? Hit reply and tell us.