Hold onto your butts, because you can now make a video from a photo right inside of Gemini.
Okay, we need to talk about Google's latest Gemini update because this one genuinely blew me away. Today Google announced that Gemini added the ability to create videos based on a static photo, powered by Veo3 and it looks like this isn't just another AI gimmick - it's legitimately impressive tech that delivers on its promises.
The concept is straightforward: upload a static photo, add a text prompt describing the scene and any audio you want, and watch as Veo 3 transforms it into an eight-second video clip, even with sound!
Simple in theory, but the execution? That's where things get interesting.
Example provided by Google
This isn't TOO surprising as Google as made this photo-to-video feature available to users in Flow, but this slide over to Gemini and expansion to even more users and even more countries shows how confident Google has become in this tech.
Google was kind enough to give us advance access and we've been playing with this feature for the past few days, and honestly, the speed of generation caught us completely off guard. We're talking and incredibly short time, 1 to 2 minutes, to generate these clips. Upload a photo of your morning coffee, describe steam rising with gentle cafe ambiance, and boom - you've got a surprisingly convincing video that brings that still moment to life.
The quality consistency has been the real standout. Unlike some AI video tools that can be hit-or-miss, Veo 3 seems to understand context remarkably well. Whether we fed it nature shots, urban scenes, or even sketches, it consistently delivered videos that felt natural rather than that uncanny valley weirdness we've come to expect from AI-generated content.
For example, I took a photo I took of the Samsung Galaxy Z Flip 7 at Samsung Galaxy Unpacked, uploaded it to Veo and entered the prompt: "Have the phone grow arms and legs and come to life and walk away" - check out the results:
Don't forget to turn up your volume!
I wanted to push the limits a little further, so I took another photo I took at Samsung Galaxy Unpacked of the exterior of the event:
And here is the prompt I entered in Gemini:
"Have the Google Android Bugdroid mascot parachute down from the sky and wave to the camera, while the people in the photo/background start walking towards the building in the background. Also please add a plane flying above with a banner attached to the rear of the plane that says "ANDROID FAITHFUL"
Check out the results below (turn up your volume):
Now clearly, this one wasn't a home run. It got the bugdroid parachuting (along with some weird dialogue of "you can do this") and while it added the plane, it didn't quite understand/get the banner concept and I have no idea what was going on with the sign on the building at the very end of the video. That said, the animating of the people of the static photo is pretty remarkable and looked totally natural.
Google's claiming over 40 million Veo 3 videos generated across Gemini and Flow in just seven weeks. That's a lot of creative experimentation happening out there. The examples they're highlighting range from reimagined fairy tales to ASMR content - because apparently someone thought "what would cutting through cooling lava sound like?" and now we can find out.
This isn't some magic trick - it's Veo 3, Google's latest video generation model that launched in May. The eight-second limit keeps things manageable while the dual watermarking system (visible and SynthID invisible) addresses the obvious concerns about AI-generated content identification.
The feature lives in both Gemini proper and Flow (Google's AI filmmaking tool), though you'll need a Google AI Ultra or Pro subscription to access it. Standard Google approach - give people a taste, then gate the good stuff behind a paywall.
Look, we've seen plenty of AI features that feel more like tech demos than actual tools. This one's different. The combination of speed, quality, and genuine creative potential makes it feel like one of those rare moments where the hype actually matches the reality.
Is it going to replace professional video production? Obviously not...yet. But I do understand how and why people in Hollywood are freaking out. This is the tip of the iceberg in terms of generative AI for video. At this moment in time though? It's a neat way to bringing static memories to life or adding that extra layer of creativity to your content.
Try it out at gemini.google.com if you've got access. Just don't blame us when you lose an hour (or until you run out of credits) turning your entire photo library into mini movies.
The photo-to-video feature is available to Google AI Ultra and Pro subscribers worldwide.
