The New York Times has entered an agreement with Amazon to license editorial content for use in the tech firm’s artificial intelligence systems. The arrangement gives Amazon access to Times material including news stories, recipes from NYT Cooking, and sports coverage from The Athletic.
According to a spokesperson for The Times, Amazon may feature this content through various customer touchpoints such as its Alexa voice service, and where practical, will provide direct links back to Times platforms for the full experience. Financial details have not been made public, but this represents the first time Amazon has signed such a content deal.
Amazon Moves Into AI Content Licensing
The New York Times has not previously entered a licensing partnership focused on generative AI, a move that follows its copyright lawsuit against OpenAI and Microsoft in which the newspaper alleged unauthorized use of its work to train algorithms. While OpenAI has finalized content deals with several other media outlets, including The Washington Post and The Guardian, this editorial content licensing agreement marks The Times’s debut in this area.
Emphasizing its ongoing commitment to protecting the value of its journalism, The Times stated that it weighs both partnerships and legal avenues as ways to manage its intellectual property rights. The partnership offers Amazon users richer news and lifestyle content, but also aligns with The Times’s commercial interests.
Meanwhile, outside the United States, attention is building around new AI models coming out of China. DeepSeek, an artificial intelligence research group, unveiled a smaller variant of its advanced R1 reasoning model by fine tuning Alibaba’s Qwen3-8B, which released in May.
Despite being more compact and less computationally intensive, DeepSeek’s new model nearly matches Microsoft’s Phi 4 and outperforms Google’s Gemini 2.5 Flash on math benchmarks like AIME 2025 and HMMT. Running the model requires a single graphics unit with higher memory, as opposed to the full size R1 which needs a cluster of GPUs.
DeepSeek has made their new model available under the MIT license, allowing both academic research and commercial deployment. The tech is already accessible through multiple platform APIs, including LM Studio, and is aimed at developers working on advanced reasoning applications while managing resource costs.
