Amazon has updated its AI video model, Nova Reel, allowing for video generation of up to two minutes in length.

The new version, Nova Reel 1.1, introduces functionalities aimed at enhancing video creation.

Launched in December 2024, Nova Reel represents Amazon’s entry into the generative video market, competing with models from companies such as OpenAI and Google.

The updated version enables users to create “multi-shot” videos with consistent styling across six-second segments, according to AWS developer advocate Elizabeth Fuentes. It now supports prompts of up to 4,000 characters.

A new feature, “Multishot Manual,” allows users to have greater control over video shots. By using a 1280 x 720-resolution image and a 512-character prompt, this feature can generate videos with up to 20 shots.

🔗 Source: TechCrunch

🧠 Food for thought

A. Business rivalry intensifies as video becomes latest frontier in AI race

Amazon’s two-minute video generation capability reflects the escalating competition among tech giants in the generative AI space, with each company pushing to differentiate their offerings.

This pattern follows a consistent trajectory in the AI market, where Amazon, Google, Microsoft, and OpenAI regularly leapfrog each other with new capabilities, driving rapid innovation across the industry1.

According to a 2024 market analysis, 22% of cloud implementations now include AI elements, showing how AI has become central to competitive positioning in cloud services2.

The Nova AI models represent Amazon’s strategic response to rivals, with the company explicitly positioning its offerings as 75% cheaper than competitors while maintaining competitive quality3.

This competitive pressure has accelerated the technical progress of AI video models, which have evolved from short clips to multi-minute content in less than a year.

B. Copyright challenges grow more complex with longer AI-generated videos

The extension of AI-generated videos to two minutes amplifies existing copyright concerns, as longer videos increase the likelihood of reproducing copyrighted elements from training data.

The U.S. Copyright Office is currently conducting a comprehensive review of how generative AI interacts with copyright law, with particular focus on training data practices that may infringe on creators’ rights4.

Recent Amazon developments