Google Is Auto-Resizing Your Performance Max Video Ads With AI: What D2C Brands Do Before Sept 4

Video is doing more of the heavy lifting in Performance Max every quarter, and most Indian D2C brands still shoot it in one shape — usually a landscape edit built for YouTube. That single orientation quietly caps your reach, because vertical placements like YouTube Shorts and in-feed slots want a 9:16 video, and if you never supplied one, your ad simply wasn't eligible there. Shooting and editing every ratio yourself is expensive, so most brands just skip it and leave that inventory on the table.

Google has decided to fill that gap for you — whether you asked or not. It's begun using generative AI to automatically create the aspect ratios your Performance Max video ads are missing, so one horizontal upload can be reshaped into vertical and square versions and run across far more placements. The feature is on by default, the rollout is already underway, and you have until September 4 to opt out if you don't want it. That makes this a decision every D2C brand running PMax needs to make on purpose, because doing nothing is itself a choice. Here's what's actually changing and how to handle it.

What Google actually changed

According to Search Engine Land, Google is now applying generative AI to resize Performance Max video ads — extending your videos into the formats needed for additional inventory. In plain terms, if your ad group has a 16:9 video but no vertical or square cut, Google's AI can generate those missing versions itself and serve them.

It's an upgrade, not a brand-new idea

Google could already do a lighter version of this: it had the ability to reframe a horizontal video into vertical and square shapes. What's new is how far the automation now reaches. Rather than just cropping or padding, the system generates the formats you never provided, treating your one upload as raw material for several. Google's stated aim, per the same reporting, is to maximise video ad reach while improving the viewing experience — which is a polite way of saying it wants your video to fill inventory you left empty.

Why this matters more for Indian D2C

Two things make this land harder here than it might elsewhere. First, mobile and vertical dominate Indian attention — Shorts, Reels-style feeds and full-screen vertical placements are where a huge share of your prospects actually watch. A brand that only ever uploaded a landscape video has been missing exactly that inventory, so on paper, auto-generated vertical cuts hand you reach you weren't getting.

Second, creative budgets are tight. Shooting a proper vertical edit of every product video costs money and time most growing D2C teams don't have to spare. If Google produces a workable 9:16 version for free, that's a genuine saving — provided the output doesn't butcher the ad. That "provided" is the whole game.

The trade-off, laid out plainly

The appeal is obvious; the catch is less so. Cropping a wide video down to a tall frame throws away the left and right of every shot, and that's where a lot of D2C creative hides its most important pixels — the price, the offer badge, the logo, the pack shot sitting off to one side. AI reframing tries to keep the subject centred, but it can't know that the ₹499 launch price tucked in the bottom-right corner is the reason the ad works.

The honest read: for a plain, well-centred product clip, leaving it on usually wins. For a dense, carefully-art-directed hero video where every corner is doing a job, an automatic crop can quietly water it down. Most brands have both kinds of asset, which is why this isn't a single toggle you flip once.

How to decide before September 4

You don't need a committee for this — you need an hour with your video library and a clear rule for each asset.

Sort your videos into "safe to reframe" and "hands off"

Pull up every video running in Performance Max and split them in two. A video is "safe to reframe" if the product and message sit in the middle of the frame and nothing important lives near the edges. It's "hands off" if it leans on corner logos, side-by-side comparisons, edge-anchored price stamps, or text that runs to the margins. Your safe list can stay on the default AI resizing; your hands-off list is where you consider opting out or supplying a proper vertical cut.

Use the opt-out deliberately, not as a blanket

Because the opt-out is account-level via Google's form or your account team, treat it as a real strategic choice rather than a reflex. If almost all your creative is simple and centred, you may not want to opt out at all. If your best-performing, highest-spend videos are the intricate ones, protecting those may be worth forfeiting some auto-generated reach. The wrong move is to ignore the September 4 date entirely and find out later that a flagship video has been running in a crop you'd never have approved.

Design future videos for the crop

The durable fix isn't fighting the automation — it's shooting so it can't hurt you. Keep your product, logo and any price or offer text inside a central "safe zone" that survives a 9:16 crop, and you get the reach of auto-resizing without the reframing damage. This kind of format-aware creative planning is exactly the sort of unglamorous groundwork our performance marketing team builds into a D2C video brief, because it's cheaper to design for every placement once than to repair cropped ads later.

Check the output, don't trust it blind

Turning the feature on — or leaving it on — doesn't end your job. Once generated versions are live, review your Performance Max asset reporting and preview the reframed formats wherever the interface lets you. The vertical version deserves the most scrutiny, since it's the one most likely to lose the sides of a landscape shot. Confirm the product is still centred, the logo isn't clipped, and any on-screen price or claim is fully inside the frame. If a generated cut looks wrong and you can't fix it in place, that asset earns a spot on your opt-out list or a purpose-shot vertical replacement.

Watch performance, not just appearance

A crop can look acceptable and still underperform, so keep an eye on how the video placements do after the change. If a previously strong video starts slipping once auto-generated formats enter the mix, that's your signal to investigate the reframed versions rather than blaming the audience or the budget. Performance Max hides a lot inside one campaign, and this is one more variable you now have to account for when a number moves.

The takeaway

Google is turning your single Performance Max video into several, using generative AI to build the vertical and square formats you never uploaded, and it's doing it by default with an opt-out deadline of September 4, 2026. For a lot of Indian D2C brands that's a quiet win — free reach on the vertical, mobile-first placements where your customers already are, without the cost of shooting every ratio. But it hands an algorithm the scissors, and a bad crop can cut the price, the logo or the message out of your best ads. So make it a decision, not a default: sort your videos into safe and hands-off, opt out where the craft matters, design your next shoot for the crop, and check the output instead of trusting it. Do that this week, and you get the reach without the reframing regret.

Sources: Search Engine Land — "Google Ads will use generative AI to resize Performance Max video ads" (Google applying generative AI to create missing aspect ratios of existing Performance Max video ads; extends an earlier capability to reframe horizontal video into vertical and square; goal to "maximize video ad reach while improving the viewing experience"; advertisers can opt out via Google's opt-out form or their account team, deadline September 4, 2026; video settings remain adjustable in-account). The India-specific application, per-asset decision framework, safe-zone creative guidance and QA steps are Digistex4u's applied recommendations.