Sonilo raises $11M to build video-native AI audio

Sonilo, a San Francisco generative-audio startup founded by veterans of TikTok's AI and music teams, announced an $11 million funding round led by B Capital with participation from Redpoint on October 14, 2026. The company said it will spend the money on compute to scale its models, on go-to-market work across creator and developer platforms, and on its music-licensing business, according to its release.

For scale, this is early-stage money. Suno raised more than $400 million at a $5.4 billion post-money valuation in June 2026, ElevenLabs raised $500 million at an $11 billion valuation in February 2026, and Synthesia, the AI video platform, raised $200 million at a $4 billion valuation in January 2026. Mirage, the company behind the Captions video app, raised $75 million in growth financing in March 2026. Sonilo's $11 million is an early-stage sum next to those rounds, and it comes with no valuation attached.

At the center of the platform is what Sonilo calls the Sound World Model, a multimodal system that connects video understanding, audio generation and temporal alignment. The company says a creator can drop in a video, describe what they want in text, or do both, and get a synchronized soundtrack. The release frames the pitch as an alternative to text-first audio tools: "Almost every audio-generation experience starts with a text prompt," said Shawn Song, Sonilo's CEO and co-founder, in the release. "We built Sonilo to understand the visual story."

Sonilo is not only selling a model. Its site lists a free tier and paid plans at $14.99 a month for Pro and $29.99 a month for Premium, as read on October 15, 2026, with commercial rights starting on Pro. It also lists plugins for Unity, Godot, Epic Games' Fab, the Roblox Asset Store, the Defold Asset Store and Adobe Exchange for Premiere Pro, plus integrations on ComfyUI, fal, WaveSpeed, Scenario and TapNow. It offers AI dubbing with synchronized lip-syncing.

The team's background is the story B Capital is buying. Song trained at Carnegie Mellon and led multimodal AI work at TikTok, per the release. CTO Alex Yin holds a PhD in computer music from the University of York and worked with Song at TikTok. COO Keli Li built and led global music at TikTok, and CMO Trista Taylor has worked on consumer AI apps that made a16z's Top 50 rankings. "We believe generative audio will become an essential part of the AI-native video stack," said Daisy Cai, a general partner at B Capital, in the release.

The competitive set is crowded and well funded, and Sonilo argues the differentiator is video-native generation, where the model reads pacing and emotion from footage rather than starting from a prompt. The video-native audio layer is where AI video vendors are already competing: ByteDance's Seedance 2.0, launched in February 2026, generates audio and video jointly, with the company's own launch post describing "a unified multimodal audio-video joint generation architecture."

Sonilo has not disclosed its valuation, its revenue or any model benchmarks, and the release gives no user or customer numbers. On the rights side, it names Shutterstock as an early licensing partner and says its models train on licensed music, with fees and a revenue share flowing to rights holders. It said it is finalizing master recording and publishing deals with major rights organizations, expected to close in the coming weeks. Variety first reported the round on October 8, 2026.

For creators, the round matters less than what it funds. More compute could mean better temporal alignment, the part that decides whether a generated score lands on a cut rather than near it. Whether Sonilo's model holds up is not something the company has published benchmarks for, so that part remains to be tested as the tools reach more users.

Sonilo plans to launch a consumer app in the fourth quarter, and the master and publishing deals it says are close will be the next public marker.