Google Unveils Gemini Omni — A Turning Point in AI Video Generation

Google released ‘Gemini Omni’ to the world. This is a shot that can change the game of AI video creation and editing competition.

What Happened

Google unveiled the first model of the Omni family, Gemini Omni Flash, at Google I/O 2026, and will be available for use in the Gemini app, Google Flow, and YouTube Shorts from May 19, 2026. That's it.

The key is 'video made through conversation'. No matter what input you put in, including text, images, audio, or existing video, you can create and edit the video by giving instructions as if chatting. The next instruction builds on top of the previous one, such as “Change this scene to night,” while maintaining character consistency. However, Google intentionally withheld some of the most controversial features in this release.

Why is it important from an ‘investment perspective’

• Reassessing Alphabet (GOOGL)’s AI competitiveness — From text (Gemini) and search to video creation, it boasted vertical integration 'from AI infrastructure to consumer apps' through its own model.

• What is decisive is that the distribution channel is directly installed in the moat — YouTube Shorts. This means that it is a company that has not only model performance, but also 'a channel that can immediately connect to billions of users.'

• Intensifying competition — The competition for AI video supremacy with OpenAI (Sora series), ByteDance (Seeddance series), etc. has intensified. There is a high possibility that the winner will take the advertising, content, and commerce markets together.

Connecting domestically,

If the unit price of AI video creation drops sharply, the cost structure of advertising, marketing, media, and commerce production will change. Even in Korea, AI content and video solutions, advertising platforms, and media production companies are directly and indirectly affected. Conversely, the ex isting video outsourcing and stock video model may be under pressure. Safety device

API·Enterprise release speed → Actual sales contribution

OpenAI·ByteDance’s tit-for-tat model performance

Advertiser’s reality Adoption rate (whether it is used to produce shorts advertisements)

Rather than the model announcement itself, the outcome of the AI video war will be determined by ‘who has a channel to connect to users faster’ It is divided.

※ This article is a summary of issues for information purposes and is not a recommendation to buy or sell a specific stock.