Bilawal Sidhu, a former member of Google’s Geo team, recently shared a demo that sparked intense discussion online. On a Google Earth screenshot, he sketched a simple flight path with his cursor. Gemini Omni, Google’s video generation model launched in May 2026, transformed that drawing into cinematic drone footage along the exact route. The demo promises a revolution in video production — but the reality is more nuanced.
Gemini Omni, also known as Omni Flash, was announced at Google I/O 2026 and began rolling out on May 19. The model can generate and edit videos based on natural language instructions and drawings. This isn’t a feature launching today, but rather a capability that has been expanding gradually to users with Google AI Plus, Pro, and Ultra subscriptions in the Gemini app and Google Flow. It’s also available free in YouTube Shorts and YouTube Create, though with different restrictions.
Drawings don’t function as a precise “drone flight path” command. Google describes this capability as a visual reference and movement guide — the camera is controlled through text, and drawings serve only to orient the generation. This means correspondence with the line you sketched is suggestive, not guaranteed to the millimeter.
There are several concrete limits. Each generated video doesn’t exceed 10 seconds, a design decision tied to computational demand. For an actual drone filming session, this means short clips, not continuous recording. Each video carries a SynthID watermark, visible in the Gemini app, the web version, and Google Search. If you intend to use these clips in commercial materials, viewers will know they are generated, not filmed.
Online discussion compared costs with traditional drone filming — an unverified estimate of approximately 1,000 dollars per session. The truth is, even if Gemini Omni makes generation more accessible, AI-generated clips and real footage accomplish different things. Generated images cannot guarantee geographic fidelity or correct architectural details, making them risky for real estate ads or materials where location accuracy matters.
Additionally, some features shown in demos — such as audio editing and voice — were still in testing at the time the model launched.
Gemini Omni genuinely has potential for conceptual exploration, rapid visualization, and creative content where accuracy isn’t paramount. But the claim that it has made drone filming “nearly free and instant” overstates what it actually does. It’s a new tool, not a complete replacement.
