Google has completed the global rollout of its cutting-edge Veo3 video generation model, making the technology available to Gemini users across more than 159 countries as of July 3, 2025. The expansion marks a significant milestone in AI-powered video creation, with Google DeepMind CEO Demis Hassabis describing it as "emerging from the silent era of video generation."
First unveiled at Google I/O in May 2025, Veo3 represents a substantial advancement over its predecessor by producing high-definition videos with fully synchronized audio. The model can generate realistic dialogue, ambient sounds, and sound effects that match perfectly with the visual content, creating remarkably lifelike results. Videos can be created in up to 4K resolution, though standard output is 720p for most users.
Access to Veo3 varies by subscription tier. Google AI Ultra subscribers ($249.99/month) receive full access to the model's capabilities, while AI Pro subscribers ($19.99/month) can generate three Veo3 Fast videos daily, each up to eight seconds in length. The technology is integrated with Flow, Google's new AI filmmaking tool that combines Veo3 with other DeepMind models like Imagen and Gemini.
The technology incorporates several safety measures to address deepfake concerns. Google's proprietary SynthID watermarking technology embeds invisible markers into generated frames with a reported 99.3 percent detection accuracy in controlled tests. Each video also contains creation metadata compliant with C2PA standards for end-to-end tracking.
Industry observers note potential applications beyond entertainment. Demis Hassabis recently hinted at possible gaming applications, responding to a social media post about "playable world models" with "now wouldn't that be something." However, the technology has already sparked controversy, with Media Matters for America reporting instances of racist and antisemitic videos generated using Veo3 appearing on TikTok in early July.