Viggle AI, a startup with a popular video generation service of the same name, today announced that it has closed a $19 million early-stage investment.

The Series A round was led by Andreessen Horowitz with participation from Two Small Fish. Toronto-based Viggle AI will invest the capital in product development and growth initiatives.

Viggle AI, officially WarpEngine Canada Inc., was founded in 2022 by Chief Executive Officer Hang Chu. He was previously a principal research scientist at Autodesk Inc.’s machine learning lab and earlier worked at Nvidia Corp. Viggle AI provides a generative artificial intelligence service that allows users to create short clips with natural language prompts.

To use the service, consumers must upload an image and a text snippet that details how the image should be animated. Alternatively, they can swap the text snippet for a video that shows the motion they wish to depict in the animation. Viggle AI uses a custom foundation model called JST-1 to automatically turn the uploaded assets into a short video.

There are several other services on the market that offer similar capabilities. According to Viggle AI, what sets JST-1 apart is that the model has a “genuine understanding of physics.” The company says that this feature allows the AI to generate more realistic character movements than competing neural networks.

Viggle AI has built up an installed base of more than four million users since making its service publicly available. Creative professionals rely on the software to create proof-of-concept animations during the prototyping phase of their projects. Hobbyists, in turn, can use it to generate clips that would otherwise require a significant time investment to create. 

The company monetizes the service through a paid Pro tier that offers higher usage limits for $9.99 per month. To grow its market presence, it will use its new funding to roll out a creator program that will give participants a Pro subscription and video creation credits. The company is also promising early access to new features. 

Viggle AI’s other priorities are to grow its headcount and enhance the AI that powers its video generation service. According to TechCrunch, the company is planning upgrades that will facilitate the creation of more sophisticated clips.

It’s competing in a crowded market. OpenAI and Runway AI Inc., which is reportedly in the process of raising a new $450 million round, also offer video generation models. Smaller contenders such as Hotshot, which debuted its newest video generation model last week, are also contributing to market competition.

Animating images is not the only use case that video generation models support. In July, Stability AI Ltd. introduced Stable Video 4D, a neural network that can take a short clip as input and automatically change the camera angle. It’s capable of recreating a video from eight new perspectives in about 40 seconds.

AI video generators are also starting to find use beyond creative workflows. Covariant Inc., a venture-backed industrial automation company, detailed a video generation model called RFM-1 in March. The model helps the company’s robotic arms plan the best way of carrying out tasks such as moving a box from one conveyor belt to another. 

Photo: Unsplash