Berlin AI startup Mirelo raises $41 million, making sound the new frontier of generative media
Berlin-based startup Mirelo, founded by former Amazon AI researchers, has raised $41 million in seed funding in one of Europe’s largest rounds in the AI-audio space. The round was co-led by Index Ventures and Andreessen Horowitz, with participation from Atlantic.vc and TriplePoint Capital.
Mirelo addresses a long-standing problem for creators: despite rapid progress in video and image generation, sound remains a bottleneck—slow, manual, and poorly synchronized with visual content. The startup’s platform uses proprietary foundation models that analyze video and automatically generate precisely synchronized soundtracks and sound effects, taking into account motion, pacing, and the emotional context of a scene.
A key technological distinction of Mirelo is efficiency. According to the company, its models require up to 50 times fewer computing resources than typical LLM systems, while delivering greater realism and sonic nuance. The latest product version, Mirelo SFX v1.5, is already available via the Mirelo Studio web app and a developer API.
The market Mirelo is entering is rapidly taking shape: in the US, ElevenLabs and Boomy focus on voice and music, while Kaiber and Runway integrate audio into video generation. Mirelo, however, is betting specifically on video-conditioned sound—audio that emerges from the image rather than being added after the fact.
The startup’s next step is interactive environments: video games, AR spaces, and dynamic worlds where sound is generated in real time and responds to user actions. For investors, this signals that sound is ceasing to be a secondary element of generative content and may become an independent infrastructure layer of the new media economy. Startup website: www.mirelo.ai
