Flam Raises $40 Million Series B to Scale AI Interactive Content Platform

AI interactive content startup Flam has raised $40 million in Series B funding, led by QED Investors, as the company looks to invest further in AI research, expand its product portfolio and grow enterprise sales internationally.

The round also included Claypond Capital, Martin Chavez, Olivier Pomel, Venky Harinarayan and Shah Rukh Khan, while existing investors RTP Global and Dovetail participated again, according to the company's official announcement.

The new round takes Flam's total equity funding to approximately $62.5 million, according to Mint. The company did not disclose its valuation for the Series B.

Flam Plans to Invest in AI Research and Global Expansion

Flam plans to put the new capital toward research and development of its AI models, expanding its product suite and scaling enterprise sales globally.

Founder and CEO Shourya Agarwal told Moneycontrol that the company intends to hire AI research engineers in the United States, expand its product team and increase its enterprise sales capacity. Flam also plans to monetize its AI models.

The expansion comes as Flam reports significant enterprise adoption. The company says it has signed more than 100 enterprise customers over the past six quarters. Its official announcement names Google, State Farm and Diageo among its customers, with applications spanning marketing, product visualization, learning and development, entertainment, customer support, fan engagement and sales.

What Does Flam Build?

Founded in 2021, Flam is developing infrastructure for interactive digital content rather than conventional one-way images and videos. The San Francisco-headquartered company has teams in India and Japan and says it holds more than 15 patents.

Its technology currently spans three main interactive formats.

Flicks are interactive videos that can change a person, product or scene while a video is playing. Flam says its patented compression technology can deliver these experiences programmatically with a 50-millisecond time-to-first-buffer. Its Forge editing model can take an image reference and modify a specified element while preserving other parts of the image.

Airboards bring high-fidelity 3D content into a camera-based interface and support touch, voice and haptic interactions. According to Flam, its 3D streaming infrastructure can load these experiences in around 300 milliseconds without requiring an app installation. The company also has an image-to-video model called Fable for creating assets as native alpha-channel videos.

The third area is Visual Agents, AI characters designed to communicate through a video-style interface rather than a conventional text chatbot.

Flam Is Also Building a 26B-Parameter AI Model

Flam's Visual Agent stack includes several internally developed technologies.

Fantom handles facial expression using identity preservation and motion transfer, while Finesse provides multilingual voice capabilities. Falcon is Flam's 26-billion-parameter mixture-of-experts (MoE) large language model.

The company claims Falcon can reach a 30-millisecond time-to-first-token, while its complete Visual Agent pipeline can return an answer in under two seconds. These are Flam's reported performance figures rather than independently benchmarked results.

This combination is intended to make interactions feel closer to a video conversation, while allowing the AI agent to both communicate with users and perform actions.

Flam Targets a Bigger International Business

Flam's ambitions extend beyond expanding its technology.

Agarwal told Mint that Flam is targeting $100 million in annual recurring revenue within 18 to 24 months, although the company has not disclosed its current ARR.

The company's revenue is currently split roughly equally between India and international markets, according to separate comments from Agarwal. Moneycontrol reported that Flam expects the US and Japan to become much larger contributors as it expands internationally.

The Series B follows a $14 million Series A in 2025. With the latest investment, Flam now has substantially more capital to develop its AI models while attempting to turn interactive content into a broader enterprise technology category.

Why This Matters

Much of the recent generative AI content market has focused on producing text, images, audio and video faster. Flam is pursuing a somewhat different direction: making generated content responsive to the person consuming it.

That could make interactive AI useful beyond advertising and entertainment. Product demonstrations could react to customer questions, training material could adapt during a session, and AI agents could combine visual communication with actions instead of relying entirely on chat interfaces.

Flam's reported enterprise adoption provides an early indication that businesses are experimenting with these formats. However, whether interactive AI content becomes a major category will depend on factors including cost, reliability, latency and whether businesses see measurable improvements over conventional video, 3D experiences and chat-based AI agents.

For now, the $40 million Series B gives Flam additional resources to test that thesis at a larger scale.