Can 3D garments and AI imagery replace e-commerce photo shoots in 2026?

KEY TAKEAWAYS Summary by the editors

1. At Zalando's Partner Day in September 2026, About You said generative AI had cut its spend on product presentation by around 90 percent, about 6 million euros a year, and WWD reported that Zalando and About You have both almost completely eliminated their dedicated content studios.
2. FashionUnited reported in March 2026 that Zalando scaled its AI-generated product content from almost zero to 90 percent within one year in 2025.
3. For most retailers in 2026, generative AI imagery rather than physics based 3D rendering is the main substitute for catalogue photo shoots; 3D remains most valuable where accuracy of fit, drape and construction matters before the garment exists.
4. H&M launched its first AI digital twin images of real models in July 2025, working with the model and photographer, which shows that consent and collaboration with talent are becoming part of the process.
5. Under the EU AI Act, providers of generative AI systems must mark outputs in a machine readable way, with a grace period to 2 December 2026 for systems on the market before 2 August 2026, and deepfakes resembling real persons must be disclosed.

Partly, and faster than many expected. In 2026 large online retailers such as Zalando and About You report replacing most studio product photography with AI-generated imagery, with About You citing a cut of around 90 percent in spend. But the substitute is mostly generative AI working from real product photos, not physics based 3D garments. 3D still matters where an image must show a product accurately before it exists, and both approaches carry accuracy, labour and legal questions.

What has changed in e-commerce imagery since 2025?

The clearest evidence comes from the large platforms. FashionUnited reported in March 2026 that Zalando scaled its AI-generated product content from almost zero to 90 percent within one year in 2025, cut campaign creation time from six weeks to a few days and increased the number of content pieces by 70 percent. At Zalando's Partner Day in September 2026, WWD reported that About You had reduced its spend on presenting products by around 90 percent by using generative AI instead of photography, saving roughly 6 million euros of a budget of about 7 million euros a year.

WWD also reported that Zalando and About You have both almost completely eliminated the dedicated studios they previously ran for content and catalogue style product pictures. According to About You's Tarek Mueller, the work began about 18 months earlier, and a breakthrough roughly six months before the event meant AI was not only reducing costs but also increasing conversion. Content featuring real talent, such as Zalando's brand ambassador, was named as one of the few areas where AI is not used.

Is this 3D or generative AI?

The distinction matters for planning. Physics based 3D garments are built from patterns and measured digital fabrics and simulate fit and drape. Generative AI imagery typically starts from photos of the real product and places it on models, poses or backgrounds that were never photographed together. The boundary is blurring as product development software adds image generation: Centric Software's AI Studio, launched in May 2026, creates commerce ready imagery connected to live PLM product data and approval workflows. For buyers, the useful question is less which technique is used than what each image is derived from, and whether that source is an approved product version.

That traceability is what separates an operational content pipeline from experiments. An image generated from a pre-production sample photo, or from a 3D file that was later changed, can show details that no longer exist in the final product.

Read also

What does 3D sampling really save? Sample costs, lead times and carbon

Where do 3D garments still make sense for imagery?

• Pre-launch and pre-order: showing styles in wholesale line sheets or on pre-order pages before production samples exist.
• Colourway multiplication: rendering every colour of a style from one 3D garment instead of photographing each.
• Technical detail views: consistent close-ups, ghost mannequin views or turntables generated from the same file.
• Brands with an existing 3D practice: where 3D files already exist from development, reusing them for imagery has a low marginal cost.

Where products are already photographed for other reasons, or where texture and drape are hard to simulate (knitwear, sheer fabrics, complex embellishment), generative approaches built on real photos are currently easier to scale.

What are the risks of replacing photo shoots?

The first risk is accuracy. An image that subtly changes a neckline, a print scale or a colour creates exactly the expectation gap that drives returns. Quality control should compare generated images with the approved product, with colour checked against physical references.

The second concerns people. H&M's approach shows one route: it announced plans for digital twins of models in March 2025 and published the first images in July 2025, working with the model and photographer and describing the process as collaborative and transparent. Rights to a model's likeness, payment and the jobs of photographers, stylists and studio teams are real issues that brands need to address openly.

The third is regulation. Under Article 50 of the EU AI Act, providers of generative AI systems must mark synthetic image outputs in a machine readable format, with a deadline of 2 December 2026 for systems already on the market before 2 August 2026, and content that appreciably resembles real persons must be disclosed as AI-generated. Brands should clarify with their tool providers how marking works and when on-site labelling is needed.

How should brands assess the return on investment?

The cost side is the easiest to measure: studio, sample, model and post-production costs per style against tool, compute and review costs. The revenue side needs controlled tests, because conversion effects reported by large platforms may not transfer to smaller brands with different traffic. A sober business case includes:

1. Current cost and lead time per style for imagery, by category.
2. Share of styles where generated imagery passes quality review without rework.
3. Conversion and return rates for AI or 3D imagery against photography in A/B tests.
4. Costs of governance: review time, rights management and labelling.
5. Effects on wholesale partners, who may have their own image requirements.

Read also

Digital showroom ROI: how to build the business case for costs and sell-in

What should brands do next?

Start with a category where products are simple and well photographed today, run generated imagery alongside existing photos, and measure accuracy, conversion and returns. Keep real photography for hero content and products where texture sells. And build the governance (consent, labelling, approval) before scaling, not after.

Frequently asked questions

Are fashion retailers replacing photo shoots with AI?

Large platforms are. At Zalando's Partner Day in September 2026, About You said generative AI had cut its product presentation spend by around 90 percent, and WWD reported that Zalando and About You have both almost completely eliminated their dedicated content studios. Real photography is still used for talent led content.

What is the difference between 3D rendering and AI-generated product images?

3D rendering simulates a garment from patterns and digital fabrics, so it can show products before they exist. Generative AI imagery usually starts from photos of the real product and creates new compositions, models or backgrounds around it.

Do AI-generated fashion images need to be labelled in the EU?

Under Article 50 of the EU AI Act, providers of generative AI systems must mark outputs in a machine readable format, with a grace period until 2 December 2026 for systems already on the market. Content that appreciably resembles real persons must be disclosed as AI-generated.

Can AI product images increase returns?

They can if they misrepresent colour, fit or details. Generated images should be reviewed against the approved product, and brands should compare return rates for AI imagery and photography before scaling.

GuideThe complete guide to AI in fashion e-commerce, marketing and retailRead the complete guide