How Inditex Spent €1.8bn on the Problem Every Mid-Market Brand Has

In 2025, Inditex — the parent company of Zara, Pull&Bear, Massimo Dutti and five other global fashion brands — invested €1.8bn in technology.

That's not marketing spend. That's not store fit-outs. That's investment in technology: AI, supply chain systems, and virtual try-on infrastructure designed to solve the most persistent problem in fashion retail.

The problem is fit. And the cost of getting it wrong at scale.

What Inditex Is Actually Solving

Inditex sells hundreds of millions of garments per year across 96 markets. At that volume, even a small improvement in return rate has enormous financial impact. A 1% reduction in returns across their online business — which accounts for roughly 30% of revenue — saves hundreds of millions of euros annually.

Their technology investment isn't charitable. It's a calculated response to the most expensive operational problem in their business: the cost of processing returned garments that didn't fit.

The technology they're investing in includes body measurement AI, virtual fitting infrastructure, and size intelligence systems that connect customer body data to specific garments. Systems that, until recently, only brands with nine-figure technology budgets could access.

The Gap in the Middle

Here's the structural problem for mid-market fashion brands.

Enterprise brands like Zara, ASOS, and H&M have the data, the engineering teams, and the capital to build or licence proprietary fit intelligence systems. They've been doing it for years.

Small DTC brands with under £500k in revenue aren't large enough for returns to be an existential problem. They lose money on each return, but volume is low enough to absorb.

The brands caught in the middle — doing £2m to £50m in online revenue, selling fitted clothing, operating on Shopify or similar — are large enough for returns to be genuinely damaging to margins, but haven't historically had access to the technology that enterprise brands use to address it.

That gap is closing.

The Democratisation of Fit Technology

The underlying models that power enterprise virtual try-on have become available via API over the last two years. The same diffusion-based inpainting technology that major fashion platforms use for garment compositing — rendering how a specific garment drapes, fits and sits on a real person's body — is now accessible to any developer.

More importantly, the body measurement layer that was historically the expensive, proprietary part of the stack — detecting shoulder width, chest, waist, hips, torso length, arm length and inseam from a single photo — can now be done in-browser using open pose estimation models, at effectively zero marginal cost.

What Inditex spent €1.8bn building the infrastructure to do, a mid-market fashion brand can now access for £99 per month.

The Same Foundation Model Class as ASOS

The try-on technology used in platforms like ASOS isn't magic. It's a specific class of generative model — trained on clothing and body data — that can realistically composite a garment onto a person's photo while preserving their pose, skin tone, body shape, and the fabric's drape properties.

That same model class is now available via commercial API. The results aren't identical to a bespoke, fine-tuned enterprise implementation, but they're close enough that the average customer can't distinguish the difference. And critically, they're good enough to materially change purchase confidence and reduce fit-related returns.

GarmCheck

UI screenshotGarmCheck try-on result — garment composited on person

What This Means for Mid-Market Brands

If you're a fashion brand doing meaningful online revenue — and you sell fitted clothing — the competitive landscape is shifting.

The brands you compete with for the same customer are investing in fit intelligence. Not all of them. Not yet. But the early movers in your segment will have lower return rates, higher conversion, and better customer measurement data feeding into their product and sizing decisions.

The technology isn't complicated to implement anymore. It installs in minutes. It doesn't require an engineering team or a technology budget. The barrier is awareness, not cost.

Inditex spent €1.8bn because at their scale, the problem is worth that investment. At your scale, the same problem is worth fixing for £99 a month.