Wikipedia (through the Wikimedia Foundation) announced a collaboration agreement with several of the largest companies in the AI ????ecosystem—including Amazon, Meta, Microsoft, and Perplexity—to use Wikimedia Enterprise, its commercial product designed to reuse Wikipedia content on a large scale in an orderly and sustainable way.

In short: if AI and "quick answer" services feed off Wikipedia, there is now a more formal way to do so, with access designed for high volume and a model that helps fund the project.

What was announced and with whom

The announcement comes as Wikipedia celebrates its 25th anniversary, and most importantly, the foundation publicly confirmed for the first time several partnerships that have been developing over the past year. Among the prominent names are Amazon, Meta, Microsoft, Mistral AI, and Perplexity, which become "clients" of Wikimedia Enterprise, the product that facilitates the reuse and distribution of Wikipedia content (and content from other Wikimedia projects) for business purposes.

It is also mentioned that Google already had a previous partnership announced in 2022, and that other agreements are also part of this package, such as Ecosia, Pleias, and ProRata (mentioned earlier), along with Nomic and Reef Media.

The key point here is that Wikipedia isn't just "letting people use it," but creating a framework to make that massive use more structured, faster, and more sustainable in a context where AI is copying, summarizing, and redistributing knowledge at scale.

Scope of the agreement: Wikimedia Enterprise and "Wikipedia at AI scale"

The heart of the agreement is Wikimedia Enterprise: a service created by the foundation to enable the commercial reuse of Wikipedia content, with optimized access for companies that need volume and speed (two things that classic scraping or informal access don't always deliver well). Simply put: if a company trains models,Whether it builds search engines or launches chatbots that answer questions based on Wikipedia,This product offers a more direct channel to obtain that information at the "pace" demanded by its systems. But the scope is not only technical; it is also strategic. The article explains that these agreements give Wikipedia another way to sustain itself in an era where much of its content is being reused by AI models and services that deliver instant answers. And that detail matters a great deal: Wikipedia is not a "free dataset that lives in the air," it is a knowledge infrastructure maintained by people, with real costs (technology, operation, community), and with a gigantic volume of use. To put it in context, the foundation notes that Wikipedia is among the 10 most visited sites in the world, with more than 65 million articles in more than 300 languages ??and nearly 15 billion views per month. In practice, that means the content that feeds half the internet is also feeding AI, and now the process of "serving" that content when used on an industrial scale is being formalized.