Google's Gemini Omni Now Makes Video Ads in Asset Studio

Google has moved another step from automated media buying into automated creative production. On August 25, 2026, Google announced that its flagship DeepMind model, Gemini Omni, now powers multimodal video creation directly inside Asset Studio in Google Ads — turning a creative brief, brand guidelines, and a website URL into campaign-ready video, free, rolling out globally.

This is a different feature from the PMax generative reframing we covered last week (that extends your existing videos into missing aspect ratios). Asset Studio with Gemini Omni starts from a brief and produces the video from scratch. Together, the two announcements mark the point where Google's AI stops being a media optimizer and becomes a video production house inside the ad platform.

What Google actually changed

The workflow is built around four steps, per Google's announcement:

1. Anchor your brand — import your existing brand guidelines and URL into Asset Studio.
2. Generate concepts — enter a brief prompt or select existing static assets to generate video storyboards and motion scenes.
3. Refine with prompts — fine-tune each scene, voiceover, pacing, and aspect ratio using natural language.
4. Package and deploy — export multi-format assets directly into Demand Gen, Performance Max, or other Google and YouTube campaigns.

Three capabilities are worth taking seriously:

• Prompt-based editing. You can refine scenes, backgrounds, or styling in real time with natural-language commands that build on previous instructions — a conversation-like workflow, not a regenerate-from-scratch loop.
• Reasoning-backed brand integrity. Google says the model doesn't just build scenes that look real; it reasons about what should happen next to keep assets consistent with your visual identity and tone.
• Multi-format scaling. Gemini Omni generates horizontal (16:9) and vertical (9:16) variations, reducing the need to produce separate versions for different placements across Google and YouTube.

The feature is free and rolling out globally. Google also shared a telling stat in the announcement: "over half of SMBs globally on Google Ads now use Google AI to create or optimize their creative" (Google internal data, Global, January–June 2026 — vendor-reported).

Why this matters for ecommerce advertisers

1. The multi-format asset gap is closing inside the platform. For years the standard advice has been: upload video in all three orientations because campaigns with more formats perform better (our PMax playbook walks through the asset-group mechanics). Gemini Omni in Asset Studio now produces 16:9 and 9:16 from one brief natively. That removes a real production burden — and it also raises the bar: if the platform can generate formats for free, the creative that wins is the one with a better message, not merely a better aspect ratio.

2. "Reasoning-backed brand integrity" is not brand approval. The model reasons about what should happen next in the context of your guidelines — but it is not your art director. A generated scene can be on-palette and on-tone while still burying a price, misreading a product detail, or softening a hard claim. The same governance discipline we laid out for PMax reframing applies here, with more force: a video created from a brief is a new version of your brand message that no human reviewed before it rendered. Review the output, give generated assets their own creative IDs in reporting, and compare them on conversions, not impressions.

3. SynthID is transparency, not quality control. Google watermarks AI-generated assets with imperceptible SynthID marking. That's good for disclosure and platform integrity — it doesn't tell you whether the creative is good. Treat it as a provenance feature, not a review process.

4. The missing piece is still performance copy. Here's the boundary worth being clear-eyed about: Gemini Omni turns a brief and brand assets into brand-consistent motion scenes. That is production. It does not ingest a product page, extract the offer, write direct-response hooks, or produce the creator-style UGC ad format that wins paid social auctions. The platform-native tools are commoditizing the render layer — the differentiation moves up the stack to script, hook, and offer strategy.

The workflow question

Every platform-native generation feature forces the same strategic question: do you supply the creative, or does the platform decide what your message looks like? Google's direction is unambiguous — from auto-generated PMax video to generative reframing to full brief-to-video creation in Asset Studio, the platform will keep absorbing basic production.

The teams that benefit most are the ones that treat platform AI as a multiplier on their own pipeline, not a replacement for it. The practical pattern: ship fresh, crop-safe, hook-first video with real product data on a cadence — the economics of that pipeline are exactly what we broke down in our AI video ads cost guide — and let Asset Studio handle the format variations of what you've already approved. When you control the message and the platform produces the versions, the automation works for you instead of around you.

If your 2026 creative plan still starts with "we need a 16:9 and a 9:16 of the same ad," that job just became free inside Google Ads. Your attention budget is better spent on the parts the platform can't do: the hook, the offer, and the UGC-style ads that make your product page feel like a real person recommending it. Generate the creative that differentiates — then let Gemini Omni make copies.

Follow-up to Google Will AI-Resize Your PMax Videos (Aug 21, 2026): that post covered generative reframing of existing PMax videos with a Sept 4 opt-out deadline; this post covers the next event in the same rollout — full brief-to-video creation in Asset Studio, announced Aug 25, 2026.