Luma Dream Machine

Luma AI, a San Francisco startup, launched Dream Machine in June 2024 as a text-to-video tool producing five-second clips. It quickly grew into a multi-model generative video product. The underlying Ray model family has iterated rapidly: Ray2 (January 2025) brought major improvements in motion physics; Ray3 (September 2025) added native 16-bit HDR and EXR export, making it the first AI video model whose output could go directly into a professional film color pipeline without conversion; Ray3 Modify (December 2025) added video editing capabilities.

By early 2026, Luma had over 25 million registered Dream Machine users and was integrated into Adobe Firefly and Amazon Bedrock (Amazon's cloud AI service). The company raised a $900 million Series C in November 2025 valuing it at roughly $4 billion, with funding tied to plans to build large-scale AI training infrastructure alongside Saudi Arabia's HUMAIN sovereign AI fund.

In practice, Dream Machine is a strong choice for cinematic mood, environmental storytelling, and image-to-video transitions. Its Ray3 model's HDR support and film-pipeline compatibility makes it the go-to for serious postproduction use cases. Luma's web interface is designed around iteration, letting you generate, compare, and refine across multiple model variants in one session, which suits creators who treat AI video as a collaborative process rather than a one-shot prompt.