What happened
Google released Veo 4 on April 15. It generates 10 to 30 second video clips from a text prompt, an image, or both. The headline capability is character consistency. The same face, the same product, the same place holds across the whole clip without breaking. Available inside Flow, Gemini, and AI Studio. Paid tier only, on Google AI plans and through the API.
Free Veo access is a separate product. Every Google account holder gets ten Veo 3.1 generations a month inside Google Vids. Veo 4 is the paid step up, and the 30-second consistent shot is what you're paying for.
What it actually does
A 30-second clip with one subject that doesn't break. That's the practical line.
For the last year, AI video has been a 5 to 10 second medium. You could generate a clip of a face, a product, or a location, but past about eight seconds the model would drift. The face would shift, the logo would soften, the bottle would lose its label. Editors worked around it by cutting fast and stitching short clips.
Veo 4 holds the subject for the full clip. That's the change. Not better motion, not better physics, not 4K resolution. A single coherent shot long enough to actually use as an ad.
The other capabilities listed in the launch (storyboarding, multimodal input, scene-to-scene continuity) are real but secondary. The single-subject 30-second shot is the thing this week.
Who should care
Anyone running a small business who wants to make a short ad around one subject.
Specifically:
- A founder doing founder-on-camera ads. Twenty to thirty seconds of you talking, generated from a single reference photo and a script. Not a replacement for a real shoot, but a viable draft surface.
- A product business with a hero SKU. A clean 15-second product spot where the bottle, the box, or the device stays on screen the whole time without warping.
- A location business that wants atmospheric B-roll of one space. The interior of a shop, the front of a studio, a single room.
Skip Veo 4 for now if your idea needs more than one main subject. Two-character dialogue, ensemble shots, a couple in frame, a team standing together. That's still where Veo breaks. The character-consistency win is per subject, not across subjects.
Also skip it if you need text on screen. AI-generated text inside video is still unreliable. Add text in post.
What to do about it
Two concrete steps.
1. Open AI Studio. Pick Veo 4. Go to aistudio.google.com. Sign in with the Google account on your paid AI plan. Pick Veo 4 from the model selector. [verify in Gemini UI on first use] If your plan doesn't include it, the upgrade prompt fires when you select it.
2. Generate one 30-second test clip of your actual subject. Not a stock idea. Your face, your product, your storefront. Write a one-line prompt. Upload one reference image. Generate. Watch the full clip and ask one question: did the subject hold for 30 seconds.
If yes, you have a new draft surface. Use it for ad concepts you would have skipped on a budget.
If no, the failure mode tells you which kind of subject Veo 4 handles and which it doesn't. That's a useful answer either way.
The honest bottom line
A 30-second coherent shot is a real production change for single-subject ads. It is not a replacement for a real camera, a real location, or a real person. It is a draft surface for the kind of ad concept that used to die in the budget conversation.
State what it is. Generate one clip. Decide from there.