Google Ads AI Image-to-Video: How Veo 3 & PMax 3.0 Are Changing Ad Creative in 2026

Image-to-video AI is no longer an experimental feature in Google Ads — it is the default creative engine for two of its most important campaign types. Starting June 30, 2026, Google's Performance Max 3.0 (fully rolled out by that date) began automatically converting product feed images and headlines into video advertisements using the Veo 2 model. Shortly after, July 1 brought mandatory creative scoring for asset groups, with low-scoring creatives automatically downweighted in ad auctions. At the same time, Google's Demand Gen format — updated in late May 2026 — now uses a Gemini-powered creative suite to generate short-form video variants from static image assets. These moves effectively eliminate the option to run image-only ads across those campaign types without manually opting out of AI video generation. For advertisers, the shift is both an opportunity to reduce production costs and a challenge to maintain brand consistency when algorithms control the output.

The Key Change: AI-Generated Video Is Now a Default Google Ads Feature

The core change is that Google Ads no longer treats video as a separate creative format that requires deliberate production. In Performance Max 3.0, the system reads product images and headlines from a merchant's feed and automatically stitches them into video ads using Google's Veo 2 model. According to a detailed analysis in Ecommerce Times, this feature defaults to "on" for new campaigns, though merchants can opt out. For existing campaigns, Google encourages adoption by linking the AI-generated videos to the campaign's performance. The rollout, completed by June 30, 2026, represents the first time image-to-video generation has been embedded natively into Google's core ad-buying system at scale.

Demand Gen followed a similar path in late May 2026. That update, covered by Online Store News, introduces a "Gemini-powered creative suite" that analyzes existing image assets — product shots, lifestyle photos, even user-generated content — and produces multiple short-form video variants optimized for different placements. The system handles variable aspect ratios, motion effects, text overlays, and music, essentially replicating what a brand's video editor would do, but at machine speed. Early benchmarks from brands testing the feature show AI-produced videos outperforming human-produced creatives in click-through rate by as much as 22% in some verticals.

How Performance Max 3.0 Uses Image-to-Video AI

Performance Max 3.0 integrates image-to-video AI in two distinct ways: automated creative generation and mandatory creative scoring. The generation side is straightforward: the system examines the product feed, selects high-quality images, and applies Veo 2's video synthesis to add motion, transitions, and text. But the scoring side may have a bigger impact on advertiser behavior. As reported in a second Ecommerce Times article (with the full URL containing a space after "d"), Google now assigns a creative score to every asset group. Low-scoring creatives are downweighted, meaning the campaign sees reduced ad delivery regardless of budget. One brand in that report shifted entirely to AI-generated video after its traditional video assets scored poorly, demonstrating how the scoring system acts as a coercive mechanism.

Additionally, the PMax 3.0 update introduces "Demand Signal Weighting," a feature detailed by Online Store News. This allows Google's AI to dynamically shift spend toward upper-funnel video inventory, driven in part by the automatically generated creatives. The result is a structural shift: brands that do not supply AI-friendly image assets may find their budgets reallocated away from lower-funnel search toward video placements they never explicitly chose.

Advertisers can still opt out of AI video generation entirely, but the downweighting of traditional creatives makes that option increasingly costly. The system is built to reward the format it best understands, which is the format it creates itself.

Demand Gen’s Gemini-Powered Video Variants

Demand Gen's approach to image-to-video is slightly different from PMax. While PMax focuses on shopping feeds, Demand Gen is designed for social-style placements across YouTube, Discover, and Gmail. The Gemini-powered creative suite in Demand Gen takes a single image asset and generates multiple video variants, each tailored to a specific placement size and context. For example, a product photo might become a 9:16 vertical video for YouTube Shorts, a 1:1 video for in-stream feeds, and a 4:5 video for Gmail promotions.

According to the Online Store News article, early testers reported CTR lifts of 22% compared to manually produced videos. That number is striking because it suggests the AI-generated videos are not merely cheaper to produce — they may actually be more effective. The likely explanation is that the AI tests multiple motion paths, image crops, and text placements, then serves the variant with the highest predicted engagement. No human production team can iterate at that scale.

Comparison: PMax 3.0 vs Demand Gen vs Manual Video Creation

The following table summarizes the key differences between the two AI-powered formats and traditional manual video creation for Google Ads.

Feature	PMax 3.0 (AI Video)	Demand Gen (AI Video)	Manual Video Creation
AI Model	Veo 2 (later Veo 3)	Gemini-powered creative suite	None
Input Source	Product feed images + headlines	Existing static image assets	Custom raw footage
Video Variants	Single automated video per asset group	Multiple size-specific variants	Typically one format per brief
Creative Scoring	Mandatory; low-score = downweighted	Not mandatory for performance, but recommended	Not scored directly by Google
Budget Optimization	Demand Signal Weighting shifts spend to video	Standard Demand Gen auto-optimization	Must be managed manually
Opt-Out Option	Yes, but penalized via scoring	Yes, with possible performance loss	Not applicable
Reported CTR Lift	Not separately reported	Up to 22% vs human-created	Baseline
Best For	Retail, ecommerce, product-specific campaigns	Brand awareness, social-style placements	Brands with strict creative control requirements

What the Benchmarks Show: AI Video Outperforms Manual in Some Cases

The 22% CTR lift reported in Demand Gen tests is the headline number, but it's not the only data point. The Ecommerce Times article on PMax noted that brands using Veo 3's video generation were able to quickly replace low-scoring traditional creatives, avoiding the delivery penalties of mandatory scoring. One brand reported that after shifting to AI-generated video, its campaign saw a 15% improvement in conversion rate within two weeks, though that figure is anecdotal.

These results align with broader trends in AI-generated advertising content. A 2026 study referenced by Searchenginejournal.com in an article by a former Google Ads strategist pointed out that AI-generated ad copy had already matched human performance, and video was the next frontier. That article, I Helped Scale Google Ads To Billions – Here’s How I’d Build An AI Search Strategy Today, argues that brands should lean into AI creative generation because the bidding and targeting systems increasingly reward ad formats that the AI can fully optimize.

Yet the benchmarks also come with caveats. The 22% CTR figure came from early tests, likely in competitive verticals where human creatives had not been rigorously A/B tested. In niches where brand identity relies heavily on specific visual treatments (luxury goods, for example), the generic motion styles produced by AI may perform worse. Google has not released aggregate benchmarks across verticals, so the numbers should be treated as directional.

The Brand Control Trade-Off: Creativity vs. Automation

The most significant concern among advertisers is loss of brand control. When an AI generates video from product images, it applies its own motion, text placement, and tone. If a brand has strict guidelines around color, imagery, and music, the AI's output may not align. The opt-out exists, but the mandatory creative scoring effectively penalizes brands that choose traditional routes. As the Online Store News article notes, this structural shift is redefining how brands allocate budgets and approach creative development. Some brands are responding by feeding higher-quality, more varied images into their feeds, essentially curating the raw material for the AI to work with. Others are investing in brand guidelines that can be translated into AI prompt parameters, though Google has not yet exposed a direct prompt interface for ad video generation.

There is also the question of differentiation. If every brand in a category uses the same Veo 2 motion models, ads may begin to look homogenous. That could hurt recall and engagement over time, potentially eroding the initial CTR gains. Google likely understands this risk and may introduce more customization options in future iterations.

How Advertisers Should Adapt Their Strategy in 2026

The practical implications are clear: advertisers should start treating their product feed images as raw material for video. Rather than relying on a separate video production pipeline, brands should ensure their image assets are high-resolution, well-lit, and varied in angle and context. The system works best when given multiple good images to choose from.

Additionally, brands should monitor their creative scores within PMax 3.0. The mandatory scoring system means that a low score directly reduces deliverability. If a brand's traditional creatives score poorly, it is better to test AI-generated alternatives than to let the campaign underperform.

For Demand Gen, the Gemini-powered creative suite offers a low-risk way to generate multiple video variants at scale. Brands should run controlled experiments comparing AI-generated videos against their best human-produced videos, using the new reporting features for new customer acquisition that Google Ads added in 2026 — as covered by Searchengineland. That reporting can help determine whether the AI videos are truly driving incremental customers or just cannibalizing existing conversions.

Finally, consider supplementing Google's native image-to-video with third-party tools for more creative control. For instance, independent offerings like the one demonstrated on imagetovideoai.tools allow you to combine multiple AI models and customize outputs. These can serve as a complement for campaigns where brand voice requires more human direction.

The era of the automatic image-to-video ad creative is here. The smartest advertisers will not fight it — they will learn to feed it the best possible ingredients and measure the results relentlessly.