Flam, the AI interactive content platform, announced a $40 million Series B led by QED Investors, with participation from Claypond Capital, Australian Gulf Capital, the SRK Family Office, Martin Chavez, Olivier Pomel and other prominent investors. Existing backers RTP Global and Dovetail doubled down. The capital will fund R&D on Flam’s proprietary AI models, expand its product suite and scale enterprise sales worldwide.
The round follows a run of growth few platforms match: more than 100 enterprise customers signed in six quarters — Google, State Farm and Diageo among them — and a clear path to $100 million in annual recurring revenue by next year. Enterprises use Flam for product visualization, brand storytelling, entertainment, fan engagement, customer support and learning.
Flam is built on a simple observation: content has evolved in waves — text, then images, then video — but every wave produced media that flowed one way, from creator to viewer. The demand for interactive content was always there. The technology wasn’t. Until now.
“Curiosity is the oldest human instinct. We don’t just want to see the world; we want to reach out and shape it,” said Shourya Agarwal, Founder of Flam. “Passive media was a limitation of technology, never of desire. With this round, we’re doubling down on the AI research that makes interactivity feel effortless, and bringing Flam to every enterprise that wants its content to invite people in rather than talk at them.”
Underpinning the platform is a technology stack protected by more than 15 patents, delivering three kinds of Flams:
Flicks — interactive videos in which generative models handle creation while AI compression algorithms optimize playback, keeping every interaction instant and smooth.
Airboards — high-fidelity 3D content streamed straight to the camera interface, generated from text and image prompts by Fable, Flam’s AI model.
Visual Agents — hyperreal avatars that hold real-time conversations as naturally as a video call. Fantom, an identity-preservation and motion-transfer model, generatively streams facial expressions, while Falcon, a 26-billion-parameter Mixture-of-Experts LLM, delivers 30-millisecond time-to-first-token inference and powers agentic actions.
Every Flam is delivered programmatically at runtime with a 300-millisecond time-to-first-buffer, and runs across any physical or digital medium.
“Interactive content has been attempted many times, and the pattern has always been the same: compelling demos that never scale,” said Nigel Morris, Co-Founder and Managing Partner at QED Investors. “What convinced us was the combination of differentiated AI infrastructure, a platform enterprises can deploy without engineering lift, and the commercial traction to prove it. We back companies that turn a hard technical problem into a repeatable business, and Flam has done exactly that.”
