The landscape for video training data and multimodal foundation models in 2026 is defined by a shift from quantity to highly ...

The video personalisation company is launching Strata, a foundation model it says produces separate, editable layers for text, animation, footage, and actors rather than a single flat file. It’s a ...

Earlier this year, Apple introduced its Foundation Models framework during WWDC 2025, which allows developers to use the company’s local AI models to power features in their applications. The company ...