Reddit has stumbled into the ultimate catch-22 of the artificial intelligence age. Shares in the discussion platform dropped by roughly 9% following reports that executives may reconsider or end Google’s access to its content for AI training.
The two companies struck a licensing deal in 2024, reportedly worth $60 million a year, allowing Google to train AI models on Reddit’s data. Impending negotiations have forced a stark calculation. Reddit’s $60 million annual payout falls short of covering the referral traffic destroyed by Google’s AI search features.
Investors read the situation the same way: if the deal that pays Reddit to train Google’s AI also enables the AI product that eats Reddit’s traffic, the economics of that arrangement requires a second look.
The Deal That Might Not Add Up
The clash between Reddit and Google illustrates a wider threat to online publishing. Traditional digital media relies on referral traffic to fuel monetisation through advertising, leads and subscriptions. AI-generated answers sever this economic cycle by answering queries before a user ever clicks through. Wealth accumulates on the platform side while publishers absorb the traffic losses.
A licensing deal changes the calculus, but only partially. The deal has to be judged against two variables at once: direct licensing revenue and indirect traffic loss. If the traffic hit is larger than the licence fee, the platform is effectively subsidising the AI product. Reddit’s reported reconsideration suggests that calculation may not be working in its favour.
This is also why the “we paid for the data” defence from AI companies only goes so far. Licensing fees for training access fail to cover the underlying migration of web traffic away from origin sites. Those are two separate economic events, and the licence fee addresses only one of them.
More from Business
- Can US Marketing Agencies Survive The Rise Of In-House AI Content?
- Who Is The Best Salesperson You’ve Never Hired?
- Are US Small Businesses Quietly Replacing Freelancers With AI Agents?
- It Costs Over £30,000 To Hire A Receptionist – Can Your Business Afford One Or Is AI The Answer?
- Did SpaceX Fly Too Close To The Sun, Or Are Investors Overreacting?
- World Youth Skills Day 2026: Are Universities Teaching Students Skills That Employers Actually Need?
- What Does A Venture Capital Analyst Do?
- The Big Disconnect Between US Federal Labor Reports And Real Life Hiring
What Are The Stakes For Everyone Else?
Reddit enters negotiations armed with top lawyers, a $60 million recurring payout and true market power. Average online publishers operate without a single one of those advantages. AI engines routinely harvest, summarise and serve their material to users. Licensing conversations don’t happen. Creators aren’t in the room where industry standards get set.
The businesses most exposed are those whose primary growth channel is organic search. SEO-driven content businesses, niche publishers, comparison sites, community platforms, service providers who built audiences through long-form helpful content. These are the kinds of businesses AI search summaries replace. They exist to answer questions that users now get answered before they click.
The problem compounds when you consider that many of these businesses have spent years producing the content that trained the AI systems now reducing their traffic. They received no licensing cheques, only the standard promise of web indexing. That same training data now powers AI tools that directly undercut their search visibility.
Why Smaller Businesses Are The Quiet Casualties Here
Treating AI training and AI search as separate economic issues is something more businesses are starting to do. Licensing content may bring in revenue for platforms with the size to negotiate it. The traffic impact from AI search is a different problem, and it’s happening regardless of whether a licensing deal exists.
The sharpest operators are adjusting their strategy by building audience channels through email lists, owned communities and paid acquisition. The rationale is quite straightforward. Channels that don’t rely on search engine goodwill remain immune when an algorithm changes its mind about sending traffic.
AI companies are building products on top of the open web while weakening the web’s traditional traffic economy. Reddit’s dispute with Google is the visible version of that tension, because Reddit has the scale to push back publicly. Smaller businesses experience that same shift without fanfare. They have neither the leverage to demand new terms nor a stock price to expose the loss.
The question the Reddit situation puts on the table for every content business is simple: are you being paid for the data, or only absorbing the damage?
