Anorak raises $5M pre-seed to make generative video playable

Generative video you can play

Anorak began as a generative video research project and now has a team of 20. It is developing proprietary inference technology intended to make generative video faster, cheaper and more controllable, so that games can move from pre-generated content to real-time generative interaction.

Its medium is interactive generative video: footage that responds to a user's choices and inputs as they happen. Rather than prompting AI versions of existing games, Anorak says it is building new play patterns and a platform from scratch, with content that is playable, watchable and shareable.

Targeting niche communities

The idea draws on what the team saw at Hutch, where car enthusiasts obsessed over detail and built communities. Anorak wants to apply that to anglers, sneakerheads, padel players, birdwatchers, K-pop fans and other groups that no publisher would greenlight a game for. Generative media, it argues, changes the economics of serving them.

The goal is for those communities to make playable content themselves. It will not be a chatbot that turns a prompt into a game; Anorak says making should be as fun as playing. Rutland argues that when making gets cheaper, value shifts to those with the taste to filter the flood.

What comes next

Anorak publishes new playable experiments on its blog every week. The funding will expand the London and Chengdu teams, bring the first playable formats to release and establish the creator platform in early 2027.