SPUR Goes Global: WAN-IFRA and FIPP Back Publisher Alliance to Shape AI Licensing Market
WAN-IFRA and FIPP join the coalition, and 30 new members sign up.
For three days in Marseille, the WAN-IFRA World News Media Congress kept circling one question: when publishers face the AI platforms, are they stronger together or alone?
On Wednesday morning the organisers gave their answer. WAN-IFRA announced a strategic partnership with SPUR, the industry coalition pushing for fair terms in AI content licensing and confirmed that 30 publishers had joined. The announcement landed moments after a panel in which one of France’s most successful AI dealmakers had spent the better part of an hour explaining why he was not signing up.
That dealmaker was Louis Dreyfus, CEO of Le Monde, welcomed pointedly from the stage as the new members were named. Le Monde has signed licensing agreements with OpenAI, Perplexity and Meta, and Dreyfus has gone on record urging other publishers to follow. His doubts about collective action were not hostile, but they were a sharp counter-argument.
The case for acting together
SPUR is the institutional bet on the opposite view. David Buttle, who founded DJB Strategies and previously ran platform strategy at the Financial Times, has been building it on a deliberately narrow premise. It is not a licensing body. It does not negotiate on anyone’s behalf. It sets the standards, transparency requirements and terms of trade that would let an AI content market function at all.
“The key relationships that matter now are with each other,” Buttle told the panel. Publishers entered the search and social eras fragmented, presented no unified demand, and lost control of their supply chains. A new distribution paradigm is forming around AI, and his argument is that the industry has one more chance to set the terms before the market hardens around it.
Doug McCabe, Chief Strategy Officer at the Guardian, made the historical version of that case the previous afternoon, and made it more forcefully. The industry, he said, is perfectly capable of collaboration when it chooses to be, and pointed to the infrastructure behind broadcast news as proof. The internet broke the habit. Each platform shift produced the same atomised scramble for traffic while the advertising economy migrated to Google and Facebook, and journalism was quietly downgraded along the way.
“What worried me most about the global internet era is that the industry as a whole has lost a kind of collective sense of agency.”
Doug McCabe, Chief Strategy Officer, The Guardian
McCabe put a number on the cost of inertia. Modelling done with Enders Analysis suggested that, over five years, a coordinated set of terms could be worth three or four times more to UK journalism than doing nothing, and that he called the conservative estimate.
The case for going it alone
Dreyfus had heard all of this before and was unmoved enough to say so. His account of the OpenAI deal was pointedly unglamorous. After ChatGPT launched, a French trade body had concluded it could not even find an address to write to. Dreyfus went to a conference, put his question to an OpenAI executive from the floor, sent an email afterwards, and closed a deal inside two months.
“You need reliable information for your LLM. We produce information. 83% of my revenues come from my readers. How can you expect me to provide reliable information without that source of revenue?”
Louis Dreyfus, President and CEO, Le Monde
The numbers he reported were the most contested moment of the session. ChatGPT, he said, was converting Le Monde subscriptions at 30 times the rate of Facebook and 173 times the rate of discovery traffic. The room did not take it on trust. Madhav Chinnappa, the session’s moderator, told him plainly that there was scepticism about the figures, and pressed on whether the underlying traffic was meaningful at all. Dreyfus conceded it was marginal and rested his case on the quality of the audience rather than its size, betting that even a halving of current conversion would leave Le Monde well ahead.
His wider point was about pragmatism. Collective bargaining, in his experience of the French negotiations with Google, moved too slowly and forced constant returns to members for sign-off. Someone large must move first and set the template. He was candid that this had made him unpopular, but unrepentant: without deals, he argued, the only alternative is litigation, and most publishers lack the resources to see a lawsuit through.
Or whether either of them is right
The most contrarian position on the panel belonged to its moderator. Madhav Chinnappa, a Reuters Institute fellow and former Google executive of 13 years, set out to disagree with everyone. His contention was that journalism’s value to the foundation models is low, because news is not central to products that are increasingly built for enterprise customers, and because the commoditisation that has dogged publishers for two decades applies here too. A single wire deal, he suggested, might give a model all the general news it needs.
Niamh Burns, senior research analyst at Enders, did not dispute that the value exists, but warned how little of the apparatus to capture it does. Advertising formats on ChatGPT shift constantly, subscription pricing is volatile, and the pot publishers are negotiating over barely exists yet.
Buttle’s toughest rebuttal was pointed at Dreyfus. A deal, he argued, is the illusion of control rather than the substance of it. Whether or not a publisher signs, its content is still taken, by the platforms directly or by third-party scrapers, and still used in ways that erode its business. The money on the table is real and taking it is rational. But the platforms are not paying for access so much as paying not to be sued. Until usage itself becomes the unit of value, with an audit trail showing how often a publisher’s work is served to a user, the market will reproduce the asymmetries of search.
The asymmetry nobody disputed
If the panel disagreed about tactics, a parallel session on collective responses left no doubt about the scale of the problem. Matt Rogerson, the FT’s director of global public policy and platform strategy, had shared a platform with McCabe the previous day. Here he walked through a regulatory picture that was improving in fragments and worsening in aggregate. The UK has paused a contested copyright opt-out, which would have let AI firms train on published work unless rightsholders actively objected, and is consulting on transparency. The CMA (the UK’s competition regulator) and the European Commission were due to set out remedies on Google’s content scraping during the congress. Brussels will not conclude its AI and copyright review until 2027.
Set against that, Rogerson argued, is a geopolitical shift that has narrowed what any regulator can do. The second Trump administration has shown it will use trade leverage to blunt international AI rules, and a UK government chasing tech investment has been making concessions it may regret. His more provocative thought was that the AI investment story may be turning: questions about energy costs, unmet promises and a tax base hollowed out by the redundancies the models are supposed to cause could yet take the shine off.
From Argentina, Martín Gonzalo Etchevers of the Argentinian News Media Association gave the asymmetry its sharpest geography. More than 600 million Spanish speakers, 19 countries, and on his account not a single comprehensive AI agreement among them. Advertising markets thinner than Europe’s or America’s, referral traffic falling away, and individual publishers with almost no leverage to bargain. His figures, that AI searches overwhelmingly end without a click and that attribution traffic has collapsed, are a trade body’s advocacy rather than independent data, but the direction is not seriously contested by anyone in the industry.
The bet, and the doubt
There is a flaw in Buttle’s model, and Dreyfus put his finger on it. Collective action rewards the small and asks the most of the large; in any such effort, as Dreyfus noted and Chinnappa agreed, the biggest players get the least back. That is precisely the calculation Dreyfus is running. He says he is watching SPUR closely, has friends among its founders and a breakfast with Buttle behind him, but cannot yet see what membership gives a publisher that already has direct relationships with the platforms and the will to use them.
Whether SPUR can answer that question will decide what it becomes. Thirty members and a WAN-IFRA partnership are a statement of intent, not a shift in power. The coalition needs the publishers who least need it, the ones with leverage to spend, and those are the publishers, like Le Monde, who have the least obvious reason to join. A coalition of the willing-but-weak would not move the platforms. A coalition that drew in the Dreyfuses of the industry might. On the evidence of Marseille, it has the argument. It does not yet have the man it most needs to convince.
This article draws on three sessions at the WAN-IFRA World News Media Congress, Marseille, June 2026.