AIWiki
Malaysia
Back to all articles
ApplicationsVeoGoogle DeepMindtext-to-video

Veo

3 min readUpdated August 2026
Veo
Type
Text-to-video generative AI model
Developer
Google DeepMind
Announced
May 2024 (Google I/O)
Latest version
Veo 3 (May 2025)
Key feature
Video generation with synchronized audio
Related
Sora, Seedance, Kling AI, Runway

Veo (also known as Google Veo) is a text-to-video model developed by Google DeepMind that generates videos based on user prompts.[1][2] It was announced at Google I/O in May 2024 and has since been updated through Veo 2 and Veo 3, the latter of which can also generate synchronized audio.[1]

History

Veo was announced at Google I/O in May 2024, with Google claiming it could generate 1080p videos over a minute long.[1] In December 2024, Google released Veo 2, which supports 4K resolution generation and has an improved understanding of physics, available through the VideoFX tool.[1] In April 2025, Veo 2 became available to advanced users on the Gemini app.[1] In May 2025, Google released Veo 3, which not only generates videos but also creates synchronized audio — including dialogue, sound effects, and ambient noise — to match the visuals, and announced Flow, a video-creation tool powered by Veo and Imagen.[1] Google DeepMind CEO Demis Hassabis described the release as the moment when AI video generation left "the era of the silent film."[1] Flow was rebranded as Google Flow at the 2026 Google I/O keynote, alongside the announcement of Google Flow Music.[1]

Key Concepts

Veo belongs to the class of diffusion-based text-to-video models, which generate moving images from text descriptions.[1] Its defining advance in later versions is native audio generation: Veo 3 produces dialogue, sound effects, and ambient noise synchronized with the visuals, addressing one of the main limitations of earlier AI video models, which produced silent footage.[1] The model is offered through Google's product surfaces, including the Gemini app and Google Flow, with Google AI credits used for access.[1]

Applications and Impact

Veo is used for content creation, advertising, marketing, and pre-production work in film and media, allowing creators to produce video from text prompts and to maintain character continuity across longer projects in tools such as Google Flow.[1] Its release has drawn attention to the societal risks of AI-generated video: in July 2025, Media Matters for America reported that racist and antisemitic videos generated with Veo 3 were being uploaded to TikTok, prompting commentary from Ars Technica on the difficulty of preventing misuse when prompts exploit the model's inability to understand subtle tropes.[1]

>See Also

References

🇲🇾Malaysian Context

Malaysian content creators, marketing agencies, and media companies use AI video generation tools such as Veo to produce social media content, product demonstrations for e-commerce, and advertising materials, competing with models such as ByteDance's Seedance and Kling AI, which are also accessible to local creators.[2] Malaysia's creative economy and digital content industry are supported by the Malaysia Digital Economy Corporation (MDEC), while the National AI Office (NAIO) has encouraged the responsible adoption of generative AI.[4][5] For Malaysian businesses, the use of AI-generated media raises questions of copyright ownership and the handling of personal data under the Personal Data Protection Act (PDPA) when generated content includes individuals.

References

  1. ↑[Veo (text-to-video model) — Wikipedia](https://en.wikipedia.org/wiki/Veo_(text-to-video_model))
  2. ↑[Veo — Google DeepMind](https://deepmind.google/models/veo/)
  3. ↑[Veo 3 — Google DeepMind](https://deepmind.google.com/models/veo/veo-3/)
  4. ↑[MDEC — Malaysia Digital Economy Corporation](https://mdec.my/)
  5. ↑[National AI Office (NAIO) — official website](https://ai.gov.my/)
Back to all articles