Sources: Reddit signed a deal with Google to make its content available for training the search giant's AI models; source: the contract is worth ~$60M per year
Social media platform Reddit has struck a deal with Google (GOOGL.O) to make its content available for training …
Context & Ripple Effects
Reddit had already been reported to have licensed its corpus to an unnamed AI company on roughly the same annualized terms. This report identifies Google as the buyer and turns a broad content-licensing signal into a concrete platform-to-model-provider arrangement.
The arrangement also supports Google’s access to Reddit’s Data API for Search, tying model training to a broader Reddit-data integration in Google Search. Later, Reddit said licensing agreements with Google and OpenAI represented about 10% of its revenue, showing that these contracts became a material business line rather than a one-off experiment.
First-order effects
- Reddit gains a reported recurring $60M annual revenue stream in exchange for making its content available to Google’s AI-model training efforts.
- Google obtains licensed access to a large body of community discussion, reducing uncertainty around using that source for training and related Search integrations.
Second-order effects
- The deal establishes a commercial benchmark for platforms with large archives of user-generated content; other AI developers may face pressure to negotiate paid, API-based access rather than rely on open-web collection.
- Reddit can use licensing proceeds to diversify beyond advertising, while Google can more closely connect a paid content input with its search-distribution advantage.
Third-order effects
- If similar agreements proliferate, high-value web content may shift from broadly crawlable material to a licensed input market, with APIs and access terms becoming strategic control points.
- The longer-term tension is that content owners may seek payment and tighter controls even as AI search changes referral economics—a conflict visible in later reports that publishers were weighing limits on Google’s AI access as traffic declined.
The trend: This is one data point in the commercialization of proprietary and community-generated web content as a controlled input for AI training and search products.