Amazon Web Services has formalized a paid partnership with Wikipedia’s parent organization under the Wikimedia Enterprise program, joining Meta, Microsoft, Mistral AI and Perplexity. Announced as part of Wikipedia’s 25th anniversary, the deal grants AWS direct API access to Wikipedia’s human-curated data, replacing web-scraping methods. Wikimedia confirmed these agreements, struck over the past year but not previously disclosed, will enable AWS to integrate verified encyclopedic content into its large language models at scale. Existing Enterprise partners include Google, Ecosia, Pleias and ProRata, underscoring a growing industry shift toward licensing structured knowledge feeds for generative AI development.

AWS has committed to extend its sovereign cloud service by adding three new EU regions—Belgium, the Netherlands and Portugal—to its existing site in eastern Germany. The expansion responds to heightened European regulatory demands for data residency and operational independence, offering customers a fully EU-based infrastructure stack. AWS’s head of European Sovereign Cloud highlighted that these additions deliver organizations more options to deploy sensitive workloads under local governance frameworks, addressing concerns over extraterritorial data access regulations such as the U.S. CLOUD Act.

AWS secured a two-year sourcing agreement with Rio Tinto to procure approximately 30,000 tonnes of copper for its U.S. data centers. The copper will be produced at Rio Tinto’s Johnson Camp mine in Arizona using the new Nuton bioleaching process, marking the first industrial-scale deployment of the technology in over a decade. Of the total volume, around 14,000 tonnes will come from Nuton-processed ore, with the remainder via conventional heap leaching. In exchange, AWS will supply cloud-based analytics to optimize Nuton operations, targeting efficiency gains in acid and water use. This partnership not only secures a low-carbon, domestic metal source but also aligns with Amazon’s broader goal of net-zero carbon by 2040 and mitigates potential supply-chain bottlenecks amid surging AI infrastructure demand.