AI video is rapidly getting more life-like with models like Veo 3 and Runway Gen 4.5, but its outputs won’t just help make movies cheaper and faster — it could lead to the creation of an entirely new art form.
That’s the perspective of Ben Horowitz, co-founder of venture capital giant Andreessen Horowitz, who recently shared his evolving view on AI video generation. Rather than seeing it as merely an efficiency tool for traditional filmmaking, Horowitz argues the technology represents something fundamentally different—a paradigm shift comparable to the leap from theater to cinema. His comments come at a moment when AI video tools are advancing rapidly, with some filmmakers already experimenting with these technologies in professional productions.
“I’m starting to realize this AI video and so forth, it’s not like making the old thing more efficient. It’s a new medium. It’s an actual new thing in the same ways that movies weren’t plays,” Horowitz said. “AI video is not video. The stories that you can tell are completely different because you can do things that without a $200 million dollar budget you had no chance of doing. And now it’s like, no problem.”
The comparison to early cinema is particularly apt. When films first emerged, many dismissed them as merely “canned theater”—a cheaper way to record stage performances. It took decades for filmmakers to realize that cinema had its own unique grammar: close-ups, cuts, montage, and camera movements that could tell stories in ways impossible on stage. Horowitz seems to be suggesting AI video will follow a similar trajectory, developing its own storytelling techniques and possibilities that we’re only beginning to glimpse.
Horowitz pointed to current adoption patterns in the film industry as evidence the shift is already underway. “The people who are on the cutting edge of the movie industry are now able to do whole movie scenes or edit or change their movie. (They can) have the AI actor do the third cut at a level of quality that even the actor doesn’t know it wasn’t them doing the acting,” he explained.
This observation touches on what may be AI video’s most immediate impact: dramatically lowering the barrier to entry for high-quality visual storytelling. What once required massive crews, expensive equipment, and months of post-production work can now potentially be accomplished by small teams—or even individuals—with access to AI tools. The democratization could be as significant as what YouTube did for video distribution or what smartphones did for photography.
“I think it’s going to change dramatically again, and there’s going to be kind of white space for not only new creatives, but new entertainment entrepreneurs that nobody is really imagining now,” Horowitz concluded.
The implications extend beyond Hollywood. Recent developments suggest Horowitz’s thesis is already playing out across the industry. OpenAI’s Sora, Google’s Veo, and startups like Runway are locked in intense competition to improve video quality and length. Meanwhile, creators are experimenting with AI-generated music videos, advertising content, and even narrative shorts that would have been prohibitively expensive just two years ago. In July this year, Netflix had said it had used AI for the first time in its Argentinian show The Eternauts, while Amazon Prime had heavily used AI in its House of David show. An Indian movie from 2012 was re-released this year with an alternative ending created by AI.
The question isn’t whether AI will change video production—that’s already happening. The question is what new artistic possibilities will emerge when creators stop thinking about AI as a tool for making traditional content more efficiently, and start exploring what becomes possible when you can conjure any image, any scene, any world from a text prompt. Just as Alfred Hitchcock couldn’t have made “Psycho” on a stage, tomorrow’s creators may produce works that couldn’t exist without AI—and we’ll need new vocabulary to describe them.