Veo 3, the game-changing video with audio generator, is shifting from pay-to-create to a free version inside YouTube Shorts.
YouTube CEO Neal Mohan announced the pending change during a speech at the Cannes Lions marketing event in France. “I am incredibly excited by the potential of AI tools to empower human creativity,” said Mohan, before revealing that Veo 3 would arrive in shorts “later this summer”.
I see 3’s inclusion follows I see 2which let creators add AI-generated backgrounds and video to their YouTube shorts. However, Veo 3 – from Google’s DeepMind division – is an AI of a different color.
Google, which also owns YouTube, introduced the generative platform in May during Google I/O as part of Gemini Ultra. It features the unprecedented ability to generate video and synced audio in one pass based on your prompt.
It’s not surprising that Veo 3 is coming to YouTube Shorts. A casual glance at rival TikTok reveals thousands of Veo 3 clips, featuring everything from fictitious (and sometimes bizarre) sitcoms and commercials to fake news reports.
If there’s been any limiting factor to Veo 3’s reach, it’s been that it’s part of the premium Gemini Ultra platform, which can cost as much as $249.99 a month. Gemini Pro ($19.99 a month) also features access, but it’s limited to one or two 8-second clips per day.
Veo 3 in YouTube Shorts will be, YouTube confirmed to me, free to use.
This is a game-changer. It means that suddenly YouTube Shorts will be the destination for generating wild, audio-filled videos, and could result in a bit of a drain from TikTok’s AI-generated activity (and its companion video editing app, CapCut).
While YouTube Shorts reportedly has more monthly active users, TikTok has the overwhelming majority of the Gen Z audience. At the very least, this update may reverse the flow of video from TikTok to Shorts and Reels, to AI video first shared on Shorts and then to TikTok and Reels.
“I believe these tools will open new creative lanes for everyone to explore,” said Mohan.
It’s what you put into it
I made this in Veo 3 as a sort of a homage to a classic Vine that I’m sure only @brittlestar, @JonathanScott, and @MrDrewScott remember. The original was funnier. pic.twitter.com/9sYbQT1Q9VJune 2, 2025
As I wrote earlier this year, the introduction of Veo 3 is a true “talkies” moment for AI video. Even though we’ve had powerful tools like Openai‘s Sora for almost a year, there’s been no way to easily realize a fully-scripted AI video with audio dialogue until Veo 3. I’ve used it to create fantastic videos featuring dinosaurs and to almost recreate the first YouTube video.
It’s not that Veo 3 will make a feature-length AI film without your input. Veo 3 prompts require script-level detail, including dialogue, scene descriptions, camera angles, shooting styles, and anything else you can think to tell Veo 3 about your production. And it should all be in the form of a prompt.
Veo 3 isn’t perfect. Videos created through the platform are usually presented in 8-second segments that can lack some continuity (I’ve noticed how the looks of the main character subtly change from scene to scene). This, though, will likely be fixed in future updates.
As for how Veo 3 will be presented in YouTube Shorts, expect it to be similar to how Veo 2 currently appears.
I have asked YouTube if there will be a limit on the number of Veo 3 videos you can generate in YouTube Shorts in a day, and will update this post with their response.