Seedance 2.0 Now Runs on US Servers via Fal

Fal says it is now hosting Dreamina's Seedance 2.0 video model on US-based infrastructure, letting American companies use the tool without inference running abroad.

Fal announced that Dreamina's Seedance 2.0 video generation model is now available to companies in the United States, with all inference handled on servers located domestically. The move addresses a common sticking point for US businesses considering foreign-developed AI models: where the actual computation happens and what jurisdiction governs that data. Seedance 2.0 has been positioned as one of the stronger video generation models currently available, and fal's pitch is that US teams can now access that capability without routing requests through infrastructure outside the country. According to a post from fal, the model also includes native audio-video generation, suggesting it can produce synchronized sound alongside video output rather than requiring a separate audio pipeline. For companies building products on top of AI video tools, hosting location has become as much a business consideration as a technical one, particularly for teams operating under contracts or compliance rules that restrict where data can be processed. By running Seedance 2.0 entirely on US infrastructure, fal is aiming to remove that barrier for American developers who want to use the model but have been hesitant about its origin or hosting setup. The announcement did not detail pricing, availability timelines, or specific enterprise customers, but it signals a broader trend of AI infrastructure providers localizing access to high-performing foreign models so they fit more easily into US-based development stacks.