Wikipedia Signs AI Licensing Deals With Microsoft, Meta, Amazon as It Turns 25
Wikipedia is monetizing AI access to its content through licensing deals with major tech firms as it celebrates 25 years of free, volunteer-driven knowledge.
Wikipedia has announced a new wave of artificial intelligence licensing agreements with major technology companies, including Microsoft, Meta Platforms, Amazon, Perplexity, and France-based Mistral AI, as the online encyclopedia marks its 25th anniversary.
The deals, revealed Thursday by the Wikimedia Foundation, allow AI companies to access Wikipedia’s vast database of human-edited articles at a scale and speed tailored for training large language models. Financial terms were not disclosed.

Wikipedia, founded in 2001, has become one of the world’s most relied-upon sources of information, hosting more than 65 million articles in over 300 languages, created and maintained by roughly 250,000 volunteers.
However, the rapid growth of generative AI has placed increasing strain on its infrastructure, as automated bots aggressively scrape content to power chatbots and AI search tools.
Founder Jimmy Wales said he supports AI companies training on Wikipedia’s content, but stressed that they should contribute to the cost of maintaining the platform.
“Our infrastructure isn’t free,” Wales said, adding that donors do not give money to subsidize large AI companies. “You should chip in and pay your fair share.”
The Wikimedia Foundation previously signed Google as an enterprise customer in 2022 and has encouraged AI developers to use its paid Wikimedia Enterprise platform rather than relying on uncontrolled scraping.
The foundation reported that while human traffic to Wikipedia has declined, automated bot traffic has surged, heavily taxing its servers.
Outgoing Wikimedia Foundation CEO Maryana Iskander said the new licensing model is designed to protect Wikipedia’s long-term sustainability without compromising its mission of free knowledge.
The agreements also aim to promote transparency in AI-generated answers by encouraging proper citation back to Wikipedia, as AI tools increasingly summarize information instead of directing users to original sources.
As Wikipedia enters its next quarter-century, the nonprofit is also exploring ways artificial intelligence could assist editors, such as fixing broken links or improving search while keeping humans firmly in control of content creation.
Despite criticism from some political figures and competition from AI-powered reference tools, Wales said Wikipedia’s human-curated model remains its greatest strength in an era increasingly shaped by machine-generated information.



