17JUN

Reach signs AWS pay-per-usage AI licensing deal

5 min read

09:26UTC

Reach plc became the first major UK regional publisher to disclose a usage-based AI licensing deal, agreeing pay-per-usage terms with Amazon Web Services as Q1 group revenue fell 6.9% and digital revenue dropped 8.1%.

← Back to Fox buys Roku's data layer for $22bn Jump to analysis ↓

IndustryDeveloping

Key takeaway

Reach's AWS pay-per-usage deal is the only AI licensing template the long tail of publishers can actually access.

Reach plc, owner of the Daily Mirror, Daily Express and more than 100 UK regional titles, disclosed in its April 2026 trading update that it had agreed an AI content licensing deal with Amazon Web Services structured on a pay-per-usage model rather than a lump sum ¹. The same statement disclosed active discussions with several other tech platforms on similar terms, alongside Q1 group revenue down 6.9% and digital revenue down 8.1% year-on-year ².

This is the structural counter to the News Corp template . Reach reaches roughly 35 million UK monthly readers across the Mirror, Express and the regional network, but carries no $1.5bn-class litigation exposure and a mid-tier balance sheet that cannot extract a one-off windfall. Pay-per-usage scales down rather than up: small payments, prospective rather than retrospective, paid against measured query volume into Amazon's Rufus assistant and adjacent products. The CJL April 2026 mapping put comparable AI licensing deals on the same scale, with DotDash Meredith at $16m a year from OpenAI, Thomson Reuters at $33m year-to-date, and Amazon paying The New York Times roughly $20m a year ³.

The model is structurally identical to the early ad-network rev-share deals that built the open programmatic web in the late 2000s, and which subsequently disintermediated those same publishers as platform integration deepened. Revenue tracks query volume rather than damages, which makes Reach's upside contingent on Amazon's product trajectory: if Rufus and adjacent surfaces grow share, Reach's revenue scales with them; if they stall, Reach is exposed to platform-side risk it does not control. The CFO disclosure of "active discussions with several other tech platforms" on similar terms suggests Reach is hedging that platform risk by spreading across counterparties before any one of them locks the rate card.

Pay-per-usage is the only template available to most of the long tail outside News Corp's tier. Reach cannot bring a $1.5bn case the way News Corp can, and Microsoft's Publisher Content Marketplace, launched February 2026, takes a 15-50% broker cut that mid-tier publishers can avoid by negotiating direct usage terms. Whether usage-based revenue offsets the 8.1% digital revenue decline disclosed in the same statement will only be testable when Reach's H1 2026 numbers land. The first publicly itemised pay-per-usage line in a published-company set of accounts will set the floor for every subsequent mid-tier negotiation in Europe.

Deep Analysis

In plain English

Reach is the company behind the Daily Mirror, the Daily Express, and over a hundred local newspapers across the UK. Like most print publishers, it has been losing advertising revenue for years as businesses shifted their spending to Google and social media. The deal with Amazon Web Services means Amazon pays Reach every time Amazon's AI systems use Reach's articles, scaled to how often those articles are actually used. The reason this is a different model from News Corp's deal is that Reach does not have the same kind of leverage. News Corp could credibly threaten to sue Anthropic for hundreds of millions and win; Reach has a bigger content archive but no equivalent legal threat. So instead of a large upfront payment, Reach gets a running payment that depends on how much Amazon's AI actually uses its content: more usage, more money. Less usage, less.

Deep Analysis

Root Causes

Reach's digital advertising revenue decline of 8.1% year-on-year reflects the structural reallocation of UK display advertising spend toward Google and Meta's walled-garden inventory. Reach cannot recover that revenue through subscription growth because its regional and national tabloid content has historically been free or near-free; the consumer price elasticity is structurally different from a premium business title like the Wall Street Journal.

The AWS deal emerges from this constraint: Reach has content volume (100+ titles, deep regional archives) but lacks the premium content scarcity that makes a lump-sum settlement like News Corp's viable.

AWS chose a pay-per-usage model because Reach's regional content is training-data utility rather than substitution-risk liability: a chatbot trained on Reach's Hull Daily Mail archive does not substitute for a paid Wall Street Journal subscription in the way Anthropic's scraping of News Corp content did. The pricing structure reflects this asymmetry in content leverage.

What could happen next?

Precedent
The pay-per-usage template, if replicated across UK regional publisher groups, creates a standardised mid-tier licensing model that small publishers can access without litigation capacity — but at rates well below the lump-sum benchmark set by News Corp at {{EVREF:event-0}}.
Short term · 0.75
Risk
Without independent audit rights, pay-per-usage models expose publishers to systematic under-reporting of consumption by AI companies — a monitoring problem that is technically and contractually complex to resolve.
Medium term · 0.7
Consequence
Microsoft's Publisher Content Marketplace, launched February 2026, is the broker version of the same template: by intermediating dozens of mid-tier publishers, it reduces the per-publisher negotiating position while capturing platform margin on the aggregate volume.
Immediate · 0.8

Source Landscape

This story draws on mixed-leaning sources from United States

United States

LeftRight

Primary parallel: The cable television retransmission consent regime in the United States (established under the 1992 Cable Television Consumer Protection and Competition Act) created a structurally identical negotiation dynamic: large broadcasters (CBS, NBC, ABC) could extract must-carry fees from cable operators, while small independent stations often settled for zero compensation.

The pay-per-usage template Reach negotiated is analogous to the retransmission consent market rather than the licence-fee market: variable, consumption-indexed, and weighted toward content that generates the most AI queries rather than toward content that generates the most journalistic value.

Counter-parallel: The music industry's transition from compulsory mechanical licensing to streaming royalty structures (Spotify's per-stream rate circa $0.003-0.005) provides the downside case.

Once a per-usage model is normalised, the price-per-unit tends to compress over time as volume scales: AI companies will argue that marginal cost of additional tokens falls as models improve, and publishers have little power to resist rate compression without a statutory floor. The News Corp settlement at {{EVREF:event-0}} is the structural alternative: a large upfront payment insulates News Corp from rate compression for the settlement's duration.

The disclosed CJL comparables bracket Reach's likely deal terms: The New York Times at approximately $20m per year from Amazon for comparable content volume (national Reach, long archive); DotDash Meredith at $16m per year from OpenAI. Reach's 100+ titles and UK-regional content archive might reasonably place its usage-based deal in the £8-15m per year range at current AI consumption rates, though no figure has been publicly disclosed.

For context: Reach's Q1 2026 digital revenue run-rate was approximately £35-40m annualised at current -8.1% trajectory. A £10m annual AI licensing contribution would represent roughly a 25% offset to the digital revenue decline.

That is commercially material for a business at this scale, but it does not reverse the structural decline; it slows it. The strategic question for Reach is whether AI licensing income buys enough Runway to complete a digital transformation before the decline outpaces the offset.

Consensus view: The Reuters Institute's Felix Simon and the Press Gazette's analysis of UK regional publisher economics both flag that Reach's -6.9% group revenue and -8.1% digital revenue in Q1 2026 create a structural dependency on AI licensing income as a counter-cyclical offset rather than a growth driver. The pay-per-usage template is, from AWS's perspective, a content-procurement model that mimics cloud metered billing: it caps AWS's upfront liability while aligning incentives with actual consumption.

Press Gazette's analysis of the CJL data places the Reach AWS deal on the lower end of known comparables: The New York Times / Amazon arrangement at approximately $20m per year, compared to Reach's portfolio reaching well over 100 titles, implies per-title value is materially lower for regional content than for branded national titles.

Counter-view: Reach's own investor communications frame the AWS deal as proof of structural viability for the pay-per-usage model in UK regional publishing: a template that works at Reach's scale (100+ titles, substantial archive) can, they argue, be replicated by smaller regional groups such as Newsquest or Archant.

The UK Press Gazette and the National Union of Journalists have separately raised concerns that usage-based licensing creates a monitoring obligation on publishers that they lack the technical infrastructure to audit: if AWS under-reports consumption, Reach has no independent verification mechanism.

Key tension: Whether Reach can independently verify AWS's self-reported usage figures, and whether the contract contains audit rights adequate to the monitoring challenge. Neither has been disclosed publicly.

Sources:Center for Journalism & Liberty / Open Markets Institute·Yahoo Finance

Mentions:Reach plc →AWS →Daily Express →DotDash Meredith →Thomson Reuters →Microsoft →Amazon →Rufus →Google →OpenAI →United States →Anthropic →Reuters →News Corp →Center for Journalism & Liberty →New York Times →

First Reported In

Update #1 · News Corp names $1.5bn AI settlement

Center for Journalism & Liberty / Open Markets Institute· 10 May 2026

Read original →

Causes and effects

Caused by

News Corp names $1.5bn Anthropic settlement

The News Corp $1.5bn settlement benchmark directly shapes what alternative AI licensing templates like Reach's pay-per-usage deal are benchmarked against

Occurred 8 May 2026

Read story →

This Event

Reach signs AWS pay-per-usage AI licensing deal

Pay-per-usage is the only AI licensing template the long tail of publishers can actually access; whether it generates material revenue at mid-tier scale will be the first hard test of the publisher-AI commercial geometry below tier one.

Led to

Alsup withholds approval on Anthropic $1.5bn

Bartz settlement floor clarifies the distinction between bilateral licensing deals (Reach/AWS, NYT/Amazon) and litigation-derived class-action floors.

Occurred 14 May 2026

Read story →

Different Perspectives

Italian market: Fastweb signed, Mediaset and RAI did not

Fastweb, the Italian broadband operator, signed Section 2 as a deployer, while Mediaset and the state broadcaster RAI are both absent. Italy therefore shows the split in miniature: a telecommunications company with one obvious deployment surface accepted the standard, and the two companies that actually broadcast to Italian audiences did not.

Paramount Skydance and its French exposure

Paramount filed a stipulation not to close on 24 July and had trial specialists admitted three days later, after nine months arguing that antitrust risk was regulatory rather than judicial. Its 22 July European clearance obliges it to divest a stake in the Universal International Pictures distribution venture and to strike no Universal distribution deal in Europe for a decade.

Time and the agent-advertising camp

Time and Mobian treat AI agent traffic as sellable inventory rather than leakage, on a site where bots now outnumber humans most days. Chief operating officer Mark Howard frames agent impressions as an extension of sponsorship, against an industry consensus still pursuing compensation for scraping.

German broadcasters

ARD, ZDF, RTL and ProSiebenSat.1 are all absent from the list, making Germany the largest single national bloc of non-signatories. Their position rests on three defensible grounds: signature is voluntary, most deployed systems are grandfathered to 2 December 2026, and a broadcaster's compliance surface spans production, archive, advertising and distribution rather than one product. None has publicly explained the decision.

Synthesia and the signed vendor layer

Synthesia signed Section 1 alongside Getty Images, Google, Meta, Microsoft, Anthropic, OpenAI, Mistral and Aleph Alpha, accepting a documented marking standard for the product they sell. London-headquartered Synthesia is the only pure AI-video vendor on the providers list, and gains a compliance artefact to put in front of European media buyers at the moment those buyers acquire a live obligation.

European Commission

The Commission published its initial signatory list on 31 July with roughly 190 organisations and let Article 50 take effect on 2 August as scheduled. Its own page stresses that adherence to the Code is voluntary while the transparency requirements are legal obligations, which frames non-signature as an evidentiary choice rather than a breach.