Speaking of powerful new innovations, Google unveiled their latest smart drone Veo 3 at the joint Google I/O 2025 developer conference. This state-of-the-art video generation model revolutionizes virtual video production. It produces top-notch visual content, purpose-built to match high-quality sound effects, background noises, and dialogue. So while Veo 3 is actually new to Gemini—but only for subscribers of Google’s AI Ultra plan, which runs $249.99 per month—through the Gemini chatbot app.
Meanwhile, Veo 3 is notable for a pair of factors, one of which is how it can be prompted with text or images. In production, users are able to interactively guide the model, directing it in the creative process of generating videos that match their vision. Depth system wide shot The system further proves a technical wizardry at detecting more advanced camera movements, including rotation, dolly and zoom camera moves. This can create a much richer, interactive, and engaging video experience.
Veo 3 does much more than simply creating visuals. It’s able to insert or remove items from video as well as recalculate the framing of shots, introducing novel functionalities. This flexibility provides users with the ability to iterate and distill their projects down to professional quality outcomes. Furthermore, Veo 3 automatically syncs generated sounds with its video clips, understanding the raw pixels of the footage to enhance overall quality.
According to Google, Veo 3 is a big improvement over Veo 2. It does wonders for the overall quality of every frame it’s able to capture. Perhaps most importantly, it adds a cutting-edge tool that allows the users to create dialogue simply by telling it what they want it to say. It’s an exciting new step forward in AI-powered content creation and video production.
The technology behind Veo 3 seems to be inspired by DeepMind’s significant prior work into “video-to-audio” AI. It is unclear what the training data source for Veo 3 was. As many point out, YouTube was a key player in its evolution.
AI video generation models have been quickly released from both startups and big tech companies. That unveiling is a sign of just how much momentum this trend is gathering. According to Veo 3, the camera is a reflection of a huge change in how content is created. It lays the foundation for audio-visual integration to be seamless and intuitive.
Demis Hassabis, co-founder of DeepMind, described the importance of this breakthrough in a 2014 interview,
“For the first time, we’re emerging from the silent era of video generation.”