6 November 2026International edition
Vol. I · No.
6 November 2026
AI IN FASHION
DAILY
The daily briefing on AI in the fashion business
Where fashion meets artificial intelligence.
Merchandising & Buying · Stories

AI fashion market grows 40.8% to $2.47bn in 2026: where the spend is going

The Business Research Company data shows spending rising from $1.75bn in 2025, but only 1% of fashion executives report mature deployments and up to 90% of pilots fail to scale.

a window display with mannequins dressed in green and black
Photo: Elena Rabkina / Unsplash

KEY TAKEAWAYS Summary by the editors

  1. The AI in fashion market reached $2.47 billion in 2026, growing 40.8% from $1.75 billion in 2025, according to The Business Research Company.
  2. Demand forecasting, visual merchandising and personalisation account for the largest shares of spending, though segment breakdowns remain unpublished.
  3. Zalando's AI assistant reached 22 million users in 2026, up from 2 million in its first two years, while H&M achieved 85 to 90% forecasting accuracy.
  4. Up to 90% of fashion AI initiatives fail to scale beyond pilot stage, with only 1% of executives reporting mature deployments despite 92% planning to increase investment.

The global AI in fashion market grew 40.8 per cent in 2026, reaching $2.47 billion, up from $1.75 billion in 2025, according to The Business Research Company. The figure covers software, services and hardware spending across demand forecasting, visual merchandising, personalisation, creative design, supply chain management and customer relationship tools. North America was the largest regional market in 2025.

Yet beneath the headline growth lies a troubling implementation gap. While 92 per cent of fashion executives plan to increase AI investment this year, only 1 per cent report having mature deployments, and up to 90 per cent of AI initiatives fail to scale beyond the pilot stage, according to research cited in McKinsey and Business of Fashion's State of Fashion 2026 report. The lag between spending and results has turned AI into a budgeting question for CFOs and merchandising directors: which applications deliver measurable return, and which categories are absorbing capital without changing the bottom line?

The question matters now because fashion companies face simultaneous cost pressure and margin erosion. McKinsey estimates that generative AI alone could add between $150 billion and $275 billion to apparel, fashion and luxury operating profits over the next three to five years, but only if companies can move AI tools from isolated pilots into daily workflows that change forecasting accuracy, inventory turns and conversion rates. For merchandising teams, that means distinguishing between AI that improves the numbers and AI that simply produces impressive demonstrations.

What does the $2.47 billion market actually include?

The Business Research Company attributes the historic growth to expansion of e-commerce fashion platforms, rising use of customer data analytics, demand for personalisation, expansion of digital retail channels and adoption of CRM tools. The applications span customer relationship management, virtual assistants, product recommendations, search and discovery, supply chain management, demand forecasting, creative design and trend forecasting.

The $2.47 billion figure represents technology spending, not operational impact. It includes software licences, cloud infrastructure, implementation services, and in some cases the hardware required to run computer-vision and recommendation engines. The Business Research Company segments the market by product category (apparel, accessories, footwear, beauty and cosmetics, jewellery), deployment mode (cloud and on-premise), components (solutions and services), application and end user (fashion designers and fashion stores). However, the firm has not published dollar breakdowns by application segment.

A separate Business Research Company report valued the visual merchandising AI market at $2.1 billion in 2026, growing from $1.71 billion in 2025 at a compound annual growth rate of 22.5 per cent. That segment focuses on automated planogram tools, in-store display optimisation and digital merchandising, suggesting overlap with the broader AI in fashion figure. AI-generated fashion photography, another sub-segment tracked by the firm, reached $2.66 billion in 2026 and is projected to grow to $8.07 billion by 2030 at a 32 per cent CAGR.

Market segment2025 value2026 valueGrowth rateSource
AI in fashion (overall)$1.75bn$2.47bn40.8%Business Research Company
Visual merchandising AI$1.71bn$2.1bn22.5%Business Research Company
AI-generated photography$2.01bn$2.66bn32.3%Business Research Company

Where is the money going in demand forecasting?

Demand forecasting represents one of the most established AI use cases in fashion, deployed by vertically integrated retailers for more than a decade. H&M uses AI-powered demand forecasting systems to predict how much of each product will sell in each store, helping the company decide what to produce and where to stock it; H&M created its AI department in 2018 and has built this capability on a Google Cloud data platform announced in 2022.

In Q3 2025, H&M's AI-driven inventory management delivered a 30 per cent profit uplift from streamlined inventory, with overstock reduced by 25 per cent and forecasting accuracy reaching 85 to 90 per cent, reducing stockouts by 14 per cent. In H&M's Q2 2026 results, the group confirmed it is using AI models to improve allocation forecasting.

Inditex invested €1.8 billion in technology and logistics for 2025 to 2026, with AI-powered demand forecasting now driving 85 per cent of initial production allocation decisions. The system ingests point-of-sale data, social media signals, search behaviour and local weather patterns to decide what to manufacture and where to distribute it. Inditex now leans on AI demand forecasting that reads signals like search behaviour, social sentiment and even local weather to decide what to make and where to send it, while RFID tags and IoT sensors give it close to full visibility of every garment from the store rail back to the design floor.

The investments produce measurable operational results. Fashion retailers that have implemented AI demand forecasting report reductions in forecast error of 30 to 50 per cent compared to traditional methods, according to practitioner analyses reviewed in 2026. McKinsey's sector-level estimates suggest AI-driven demand planning and inventory optimisation can deliver 20 to 50 per cent lower inventories and 20 to 50 per cent fewer forecast errors.

How much is being spent on visual merchandising AI?

The visual merchandising artificial intelligence market grew from $1.71 billion in 2025 to $2.1 billion in 2026 at a compound annual growth rate of 22.5 per cent. The segment includes automated planogram generation, shelf optimisation, digital display management and product placement analytics. Retail stores represent the largest application segment for visual merchandising AI, accounting for 38.4 per cent of the market and including specialty retailers, department stores and category-specific stores that depend on visual merchandising excellence for sales conversion.

Key companies operating in the visual merchandising artificial intelligence market are focusing on developing advanced AI-powered tools, such as store-specific AI planogram tools, to enhance efficiency, accuracy and sustainability in retail operations; store-specific AI planogram tools refer to advanced software platforms that automatically create tailored visual display plans for individual retail locations using real-time inventory data and business objectives; it is an AI-driven visual merchandising platform featuring a generative AI engine that auto-generates custom planograms for individual stores in one click, dramatically reducing the time and complexity of manual merchandising.

Several European fashion retailers integrated AI outfit generators and visual merchandising automation, cutting campaign lead times from weeks to hours. The speed gains come from automating layout decisions that previously required manual coordination across buying, merchandising and creative teams. However, the category sits at the intersection of creative judgement and data-driven placement, and results depend on how well AI recommendations align with brand aesthetic and local customer behaviour.

Only 1% of fashion executives report mature AI deployments
1%Share of fashion executives with mature AI implementations in 2026

Source: McKinsey, State of Fashion 2026

What is personalisation spending achieving?

Personalisation spans product recommendation engines, dynamic pricing, AI-powered search and customer-facing assistants. The category absorbs a significant share of AI budgets because it touches every digital interaction, from homepage merchandising to checkout. Industry analyses suggest fashion e-commerce platforms using advanced AI personalisation report 15 to 35 per cent increases in click-through rates, 10 to 20 per cent improvements in conversion rates and 20 to 40 per cent increases in average order value compared to non-personalised experiences.

Zalando's enhanced AI Assistant drove a 63 per cent increase in high-value actions, defined as add-to-wishlist, add-to-cart and purchase activities, between January and June 2026. By prioritising products available in customers' sizes, the Assistant has already increased add-to-cart rates by 4 per cent in A/B testing. 22 million people visited the Zalando Assistant in 2026, up from just over 2 million in the first two years; the assistant is available in all 29 Zalando markets, in local languages, on app and web.

Interactions with the assistant last three times longer than conventional searches and average more than two questions per conversation, according to data provided by Zalando. Zalando scaled AI-generated marketing and product content from near zero to 90 per cent of its total output in a single year, compressed campaign creation from six weeks to a matter of days, and by early 2026 the pipeline had produced 170,000 AI-generated product videos.

Personalisation ROI depends on basket size, conversion lift and customer lifetime value. The challenge for merchandising teams is isolating AI contribution from other variables. A/B tests provide the cleanest measurement, but they require sufficient traffic and disciplined holdout groups. When personalisation is deployed across an entire platform, the counterfactual disappears, and it becomes difficult to separate AI impact from seasonal trends, pricing changes or assortment decisions.

How are fashion companies addressing fit and returns?

Size recommendation and fit prediction represent a distinct spending category driven by the economics of returns. Fashion e-commerce brands that deploy AI size recommendation report 10 to 20 per cent reductions in size-related returns, directly improving margin, according to industry case studies compiled in 2026. The segment includes body-measurement tools, virtual try-on, fit advisors and returns analytics.

Documented case examples show variable results. One direct-to-consumer fashion brand using an AI size recommendation tool cut returns by 22 per cent, lifted average order value by 15 per cent and improved product page conversion by 15 percentage points within four months. A fair-fashion brand deploying a fit finder achieved a 29.9 per cent conversion lift, a 2.4 per cent reduction in returns and a 47 per cent increase in average order value in three months. A premium denim case reported a 45 per cent reduction in overall returns, a 62 per cent cut in size-related returns, a 28 per cent uplift in conversion and a 19 per cent increase in average order value.

However, the evidence is not uniform. A 2025 observational study using a large fashion-platform dataset found that size-finder use was associated with a small increase in returns while also being associated with higher later customer value. The discrepancy suggests that implementation details, data quality and integration with existing workflows determine whether fit AI delivers net benefit. Some fit-recommendation spending has gone to tools that promised 30 to 40 per cent reductions in returns but delivered 2 to 8 per cent in practice, according to operator audits.

AI ROI by use case in fashion operations
Reported improvement range (low)Reported improvement range (high)
Forecasting accuracy
20 % improvement
50 % improvement
Inventory reduction
20 % improvement
50 % improvement
Conversion lift (personalisation)
10 % improvement
35 % improvement
Return reduction (fit AI)
2 % improvement
22 % improvement

Source: McKinsey, Zalando, H&M, case studies compiled 2026

Which companies are deploying AI at scale?

Scale separates pilots from production systems. LVMH has a companywide generative AI agent called MaIA that is handling more than 2 million requests a month from 40,000 employees. LVMH brands are using AI tools to summarise previous interactions with customers and help sales advisers generate personalised messages; better understand what customers are searching for on the brands' websites; adjust prices based on currency fluctuations and other factors; identify and react more quickly to supply chain problems; construct mood boards to inspire design teams; and generate copy for e-commerce sites.

To power its AI tools, LVMH built a central data platform in collaboration with Google over the past four years. LVMH has committed over €100 million to AI, expanded its Google Cloud partnership in 2025, and runs an internal assistant, MaIA, that handles more than two million requests a month. The company showcased AI applications at VivaTech 2026, including an AI assistant for Celine powered by MaIA, AI-enhanced leather-cutting for Louis Vuitton artisans and generative engine optimisation for Dior.

Zalando provides a detailed public example. Nearly 10 million customers interacted with the Zalando Assistant in Q1 2026, compared to 6 million during the whole of 2025. Zalando reported group GMV of €4,915 million in Q2 2026, representing a 20.7 per cent year-over-year increase, while revenue climbed 20.8 per cent to €3,424 million. The AI programme delivered measurable results: delivery-promise precision improved by 22 percentage points, a technology organisation of roughly 3,000 engineers shipped over 20 per cent more code changes, the Size & Fit feature avoided more than 8 per cent of size-related returns, and foundational-model matchmaking lifted items added to bags by 13 per cent.

H&M built a 270-person AI unit and deployed forecasting, inventory allocation and trend detection tools across its global operations. Inditex operates what analysts describe as one of the most sophisticated demand-forecasting AI systems in the apparel industry, dating back over a decade and continuously enhanced. The common factor is not the size of the AI budget but the integration of AI outputs into daily decisions about assortment, pricing, allocation and replenishment.

Key AI deployment milestones in fashion, 2025 to 2026
  1. 2018H&M establishes dedicated AI department, begins building demand forecasting on Google Cloud
  2. Early 2024LVMH launches MaIA, internal AI assistant, begins Google Cloud data platform collaboration
  3. 2025Zalando scales AI-generated content from near zero to 90% of marketing and product output
  4. Q3 2025H&M reports 85-90% forecasting accuracy, 25% overstock reduction, 30% profit uplift from AI inventory tools
  5. Q1 2026Zalando Assistant reaches 10 million users; LVMH's MaIA handles 2 million monthly requests
  6. 2026Zalando Assistant expands to 22 million users; Inditex AI drives 85% of production allocation decisions

Source: Company reports, McKinsey, Business Research Company

Why do most AI pilots fail to reach production?

McKinsey and Business of Fashion's State of Fashion 2026 says research suggests up to 90 per cent of initiatives fail to scale beyond the pilot phase. This year, 92 per cent of fashion executives intend to increase their AI investment, yet only 1 per cent report having mature deployments; on top of that, 90 per cent of pilots fail to scale. MIT's 2025 study on the GenAI Divide found that just 5 per cent of integrated AI pilots are extracting millions in value, a finding Fortune summed up as 95 per cent of pilots failing.

The gap between pilot success and production failure stems from evaluation criteria that do not reflect operational reality. A pilot that generates impressive product images or accurate size recommendations in a controlled test often breaks when exposed to incomplete data, inconsistent SKU hierarchies, legacy system constraints or the need for human validation at scale. Fashion AI pilots typically measure model accuracy or creative output quality, but production deployment depends on data pipelines, API reliability, change-management training and integration with PLM, ERP and WMS systems.

Three failure modes recur across fashion AI implementations. First, the pilot uses clean, curated data, but production systems contain incomplete attributes, inconsistent taxonomies and manual overrides that the AI cannot parse. Second, the pilot runs in a vendor demo environment, but production requires integration with SAP, Oracle or proprietary systems that were not designed for real-time AI queries. Third, the pilot demonstrates technical feasibility, but the organisation lacks the processes to route AI recommendations into buying, merchandising or design workflows where they can change decisions.

Cost and ROI ambiguity also contribute to stalled deployments. AI-powered demand forecasting solutions for mid-market fashion companies start at €30,000 to €80,000 per year on a SaaS model, according to operator estimates compiled in 2026. Enterprise implementations involving custom models, data integration and change management can exceed €200,000 to €750,000. Fashion executives struggle to build a clear business case when the ROI depends on variables such as markdown avoidance, stockout reduction or conversion lift that are difficult to attribute solely to AI.

What are the risks and limits of AI in merchandising?

AI in fashion carries operational, financial and reputational risks that become visible only at scale. Demand forecasting models trained on historical sales data can amplify bias, over-index on bestsellers and under-represent emerging trends or niche categories. If the model recommends conservative assortments, the business may miss margin opportunities in higher-risk, higher-reward product. If it recommends aggressive buys based on social signals, the business may accumulate dead stock when trends reverse faster than the supply chain can respond.

Visual merchandising AI introduces a different set of concerns. Automated planograms optimise for sales velocity, but they may deprioritise brand storytelling, seasonal theming or the kind of editorial curation that differentiates premium retailers. AI-generated product imagery can reduce photography costs by 90 per cent and compress timelines from weeks to days, but it also raises transparency questions. Levi's deployment of AI-generated models in 2025 attracted criticism over diversity representation and the displacement of human models, illustrating the reputational risk when AI touches customer-facing creative.

Personalisation at scale depends on customer data, and fashion companies must navigate privacy regulation, consent management and the risk of algorithmic filter bubbles that limit product discovery. AI shopping assistants and recommendation engines can increase conversion, but they can also create a feedback loop in which customers see only the products they are statistically likely to buy, reducing exposure to new styles, designers or categories. For merchandisers, the challenge is balancing short-term conversion gains with long-term customer engagement and brand differentiation.

The technology itself remains imperfect. Fit recommendation tools that rely on self-reported measurements or return-history data can produce inconsistent results when garment specifications are incomplete or when customers order multiple sizes with the intention of returning most. Virtual try-on tools can demonstrate how a garment drapes on a standard avatar, but they struggle with complex fabrics, tailored fits and the subjective comfort factors that drive keep-or-return decisions. These limits do not make the technology unusable, but they require realistic expectations and contingency processes.

Finally, AI dependencies introduce fragility. If a demand-forecasting model fails during peak buying season, or if an API outage disables personalised search during a sale event, the business can lose revenue and customer trust. Fashion companies deploying AI at scale need fallback systems, monitoring dashboards and incident-response protocols that treat AI tools as mission-critical infrastructure rather than experimental projects.

What are the most common questions about AI fashion spending?

What should fashion companies do now?

  1. Measure operational outcomes, not demo quality. Define AI success by changes in forecast error, markdown rate, conversion or return rate, not by how impressive the pilot output looks. Establish baseline metrics before deployment and track them weekly.
  2. Map AI outputs to actual decisions. Identify which person or system will consume the AI recommendation and what action they will take. If the demand forecast does not change the purchase order, or the size recommendation does not alter the customer's selection, the AI adds cost without value.
  3. Audit data quality before model selection. AI performance depends on clean, consistent, complete product data. Conduct a data-readiness assessment covering SKU attributes, size charts, historical sales, customer behaviour and return reasons before committing budget to models or vendors.
  4. Build integration into the pilot scope. Treat API design, system handoffs and workflow changes as pilot deliverables, not post-pilot tasks. If the pilot cannot demonstrate that AI outputs flow into SAP, Oracle, Shopify or the internal PLM system, it has not proven production readiness.
  5. Start with high-ROI, low-complexity use cases. Demand forecasting, size recommendation and automated content generation deliver measurable return and require fewer workflow changes than creative design or trend prediction. Prioritise applications where the business case is clear and the data exists.
  6. Establish AI governance and fallback protocols. Define who monitors AI performance, who intervenes when outputs are wrong, and what happens if the AI system fails. Treat AI tools as mission-critical infrastructure with incident response, change logs and version control.
  7. Invest in training, not just technology. AI changes how buyers, merchandisers and designers do their jobs. Allocate budget for onboarding, workflow redesign and ongoing support. MIT research shows that only 5 per cent of AI deployments extract millions in value, and execution capability is the differentiator.

Frequently asked questions

How large is the AI in fashion market in 2026?

The global AI in fashion market reached $2.47 billion in 2026, growing 40.8 per cent from $1.75 billion in 2025, according to The Business Research Company. The figure covers software, services and hardware spending across demand forecasting, visual merchandising, personalisation, supply chain management and creative tools. North America was the largest regional market in 2025.

Which fashion AI applications deliver the clearest return on investment?

Demand forecasting, size recommendation and automated content generation show the most measurable ROI in 2026. H&M achieved 85 to 90 per cent forecasting accuracy and a 30 per cent profit uplift from AI inventory management. Zalando's AI Assistant drove a 63 per cent increase in high-value actions and reached 22 million users. Fit recommendation case studies report 10 to 22 per cent return reductions and 15 to 47 per cent increases in average order value, though results vary by implementation quality.

Why do 90 per cent of fashion AI pilots fail to reach production?

Pilots fail to scale because they use clean, curated data and controlled environments, but production systems contain incomplete attributes, inconsistent taxonomies and legacy constraints. McKinsey research cited in State of Fashion 2026 found that up to 90 per cent of AI initiatives fail to scale beyond pilot stage. Only 1 per cent of fashion executives report mature AI deployments, despite 92 per cent planning to increase investment. Successful deployment requires data quality, system integration and workflow redesign, not just model accuracy.

Stories