
The AI arms race intensifies, not only because models become better, but because they become more disruptive.
What began with the launch of OpenAI’s ChatGPT, which redefined how people interact with machines, it opened the floodgates for a new wave of creative possibilities. Google quickly followed with Gemini, a suite of advanced models designed to rival ChatGPT, showcasing the tech giant’s commitment to pushing the boundaries of artificial intelligence.
But AI innovation didn’t stop at text.
One after another, Google's DeepMind then introduced Veo 3, a cutting-edge AI video generation model that reimagined how creators could produce video content.
Originally unveiled at Google I/O 2024, Veo 3 was designed to compete with OpenAI’s Sora and other emerging video-generation models like Pika, and Runway, and a handful of others.
After first announcing its plan, YouTube finally makes this very tool available on its platform, allowing creators to use Veo 3 to create Shorts.
Veo 3 is rolling out in @YouTube Shorts! Explore some of the new AI features we highlighted at #MadeOnYouTube ↓ https://t.co/DVuKZCpODI
— Google (@Google) September 16, 2025
Veo 3 allowed creators to turn text prompts into highly polished video content, complete with synchronized audio. One of its standout features was the ability to edit and remix previously generated footage, giving creators a level of flexibility and control never seen in AI video tools before.
With the model arriving to YouTube, it represents a major evolution from YouTube’s earlier Dream Screen feature, launched in 2023, which allowed users to generate AI-powered backgrounds for Shorts using text prompts.
Veo 3 brought this capability to the next level, enabling six-second standalone clips for the first time and offering creators more sophisticated tools to enhance storytelling with filler scenes, smooth transitions, and cinematic flair.
The announcement of Veo 3's arrival to YouTube was made at the 2025 Made on YouTube event, where YouTube said that creators can simply select “Create”, enter a prompt, and have the Dream Screen generates a selection of images, one of which can then be transformed into a video.
The creations are watermarked with DeepMind’s SynthID tech to indicate AI-generated content.
But unlike the original Veo 3, which can generate 1080p cinematic clips, this YouTube Shorts version, dubbed 'Veo 3 Fast' for Shorts, allows creators to only produce vertical clips at 480p resolution for rapid generation, while retaining cinematic styles and synchronized sound.
Beyond visuals, the AI can also help brainstorm video ideas, produce thumbnails, and even draft AI-assisted replies to followers within YouTube Studio, streamlining content creation end-to-end.
YouTube is launching this Veo 3 Fast to users in the U.S., UK, Canada, Australia, and New Zealand, with plans to expand to more soon.
Veo 3’s arrival on Shorts is part of a broader push to enhance creator engagement and monetization.
The goal is to creators’ imagination at the forefront.
Soon, users should be able to add motion, bringing photos to life by transferring movements from other videos, whether it’s a fun dance or a dramatic sports move.
Creators can also stylize their videos, applying artistic effects like pop art or origami with just a tap, and add objects by describing them in text, from a whimsical rubber ducky to a giant octopus, creating limitless possibilities for storytelling.
With Veo, you can also make your photos dance.
Under the hood, Add Motion uses a detection model to capture and transfer movement from one subject to another pic.twitter.com/8xFMUp4k3q— Google DeepMind (@GoogleDeepMind) September 16, 2025
Beyond these creative tools, YouTube is also rolling out 'Edit with AI,' which helps creators start videos from raw footage.
This feature intelligently assembles the best moments, adds music, transitions, and playful voiceovers in English or Hindi, providing a polished first draft so users can jump straight into personalizing their content.
Edit with AI can help take your raw footage and turn it into a compelling first draft with music , transitions , and even a voiceover .
Dropping on @YouTube Shorts and the YouTube Create app later this year. pic.twitter.com/UF36IrrFvF— Google DeepMind (@GoogleDeepMind) September 16, 2025
Another addition is 'Speech to Song,' which allows creators to turn dialogue from eligible videos into catchy soundtracks using Lyria 2, Google DeepMind’s advanced AI music model. Creators can add their own vibe, like chill, danceable, or fun, while ensuring the original creator is credited.
Implications for creators is significant.
By integrating Veo 3 into Shorts, YouTube democratizes video creation, enabling anyone with a smartphone to produce professional-quality content. From cinematic filler clips to full AI-generated backgrounds, creators can elevate the storytelling quality of their Shorts without extensive equipment or editing expertise. With over 200 billion daily Shorts views, YouTube is solidifying its position as a short-form video powerhouse, while also empowering creators to harness AI to its fullest potential.
The integration of Veo 3 into YouTube Shorts represents a significant milestone in AI-driven content creation. From the LLM war sparked by ChatGPT to the arrival of Gemini and now Veo 3, AI is redefining how we produce and experience media.
YouTube’s adoption of Veo 3 Fast ensures that creators have the tools to transform imaginative ideas into polished, shareable content, shaping the future of short-form video creation for years to come.