Publishers built the meter for AI licensing. No AI company has plugged it in.

What happened

A publisher coalition called SPUR — Standards for Publisher Usage Rights — released an open-source telemetry standard for tracking what AI systems actually do with content after they fetch it.

Founding members include the BBC, Financial Times, Guardian Media Group, Mediahuis, Sky News and Telegraph Media Group. The Associated Press joined as the first US founding member.

The spec is published under Apache 2.0 and tracks five events: retrieved, grounded, cited, presented, engaged.

One caveat on timing, and it's unresolved. The standard's own documentation dates v1.0 to 2 September 2026. Trade coverage describes the release as 2 October. Both are on the record. Nobody has reconciled them.

The gap it's built to expose

Right now a publisher sees exactly one thing: an HTTP request hitting the server. Something fetched the page. That's it.

What happens next is invisible. Did the content enter the model's generation context? Was it cited? Was the citation ever made visible to a human? Did anyone click?

SPUR names three failure modes directly: retrieval without grounding, grounding without citation, citation without presentation.

"Grounding without citation" is the technical name for your work shaping the answer and you getting nothing. That's the whole fight, in four words.

Characters, not tokens

The spec designates character counts — chars_ingested and excerpt_chars — as the portable units. Token counts get pushed to supplementary status.

That is not a footnote. Tokens vary by whose tokenizer reads the passage. The vendor controls the tokenizer, so the vendor controls the count. Characters don't move.

Publishers picked the unit of account the counterparty can't adjust. If you ever negotiate a usage-based AI deal, this is the single most useful thing in the document.

The catch

No AI company has accepted SPUR's invitation to its advisory board. Google gave a non-answer. The others did not respond.

The standard is voluntary. There is no enforcement mechanism, no regulator behind it, and no established adoption. A publisher can emit these events; nobody can compel an AI company to report back.

What this is NOT

This is not a payment system. There is no rate card, no dollar figure, and no publisher has disclosed a deal priced using this spec.

It is not a blocking tool. It doesn't stop crawlers. Cloudflare, robots.txt and Google-Extended still do that job.

It is not a video or audio standard. SPUR is built by text news publishers for text news. If you make YouTube videos or podcasts, there's no equivalent meter for you — and that's the point worth noticing.

And it's not evidence that anything is being measured today. It's a proposal for how measurement would work if the other side agreed.

Why it still matters to you

The work here isn't the launch. It's the vocabulary. Five events give you words for a harm that previously had none, and a licensing negotiation runs on defined terms.

The process was not casual: 171 test fixtures, 15 validated worked examples, 59 mutation checks in continuous integration. Public comment ran 12 June to 24 July 2026 and produced 29 specification threads, 3 profile threads and 5 pull requests.

What to do about it

1. Read the five event names and use them. "Retrieved, grounded, cited, presented, engaged" is now the shared language. Showing up to a conversation without it costs you.
2. Put character counts in any AI usage term sheet. Not tokens. If a counterparty insists on tokens, ask whose tokenizer and why it matters to them.
3. Add one clause to your next licensing draft: usage reporting at the grounded stage, not just retrieval. That's where the current blind spot is.
4. Do not plan revenue around this. Zero AI companies have signed on. It's a measuring stick, not a check.
5. If you publish video or audio, ask your trade body what the equivalent looks like. Right now there isn't one.
6. Track the advisory board roster, not the press releases. The day an AI company joins is the day this becomes a story about money.