In a landmark strategic move announced on January 15, 2025, the Wikimedia Foundation has unveiled a series of commercial partnerships with leading artificial intelligence companies, including Amazon, Meta, Microsoft, Mistral AI, and Perplexity AI. These deals are facilitated through the foundation's commercial arm, Wikimedia Enterprise, a dedicated API product designed for large-scale, high-volume reuse of content from Wikipedia and its sister projects.
The newly confirmed partners join existing client Google, with the foundation also noting ongoing relationships with entities like the search engine Ecosia, and firms such as Pleias, ProRata, Nomic, and Reef Media. This broad client list underscores the universal demand for Wikipedia's trusted data. Selena Deckelmann, Chief Product and Technology Officer at the Wikimedia Foundation, emphasized the initiative's importance, stating: "Wikipedia shows that knowledge is human, and knowledge needs humans. Especially now, in the age of AI, we need the human-powered knowledge of Wikipedia more than ever."
The partnerships address a critical need for AI companies that require reliable, structured, and fast access to verified factual data for training large language models (LLMs) and powering real-time applications like chatbots and search engines. Before Wikimedia Enterprise, companies typically scraped Wikipedia's public website, a process that was inefficient and strained the site's servers. The commercial product offers a dedicated pipeline with clear licensing, volume tiers, and professional support.
The announcement was made as part of Wikipedia's 25th-anniversary celebrations, which also included a birthday campaign, an interactive time capsule, and a livestreamed event. The timing is significant as the foundation confronts shifting traffic patterns; it reported in October that human visits to Wikipedia fell 8% year-over-year, attributing the decline to users relying on AI-generated summaries. The revenue from these partnerships provides a crucial, diversified income stream to sustain the nonprofit, which relies primarily on donations, to support its massive infrastructure serving nearly 15 billion page views per month across over 300 languages.
The move sets a powerful precedent in the ongoing debate over AI training data, demonstrating a formal model for collaboration that mitigates legal and ethical risks associated with unsanctioned data scraping. This comes amid heightened regulatory scrutiny and lawsuits from publishers and rights holders against AI firms for alleged copyright infringement.
