The first visual for a new garment is a sketch or a mood board. The last document before it hits the factory floor is a technical pack, a dense specification listing every measurement, stitch, and material. The gap between those two artifacts is where Fermat, a Barcelona-based startup, is building its business.
Founded in 2022, Fermat sells AI software that aims to compress that gap. Its platform takes a designer’s initial concept and generates realistic product images, allowing for rapid iteration on materials and styles. The more consequential step is its Techpacks product, which converts those approved designs into the factory-ready documents that guide physical production [Sesamers, November 2026]. The company recently raised an $8 million seed round to expand this workflow, particularly into the US market [FashionNetwork, November 2026].
The Wedge: From Visuals to Vendor Specs
Generative AI for image creation is a crowded field. Fermat’s initial differentiation is training its models specifically on fashion garments and materials, which the company says leads to more usable and realistic outputs for designers experimenting with cuts, colors, and fabrics [Pomegra, November 2026]. The real wedge, however, is the downstream workflow. A visual render is not a production instruction.
Techpacks automate the creation of technical specification sheets. This involves translating a visual design into precise technical flats, detailed construction notes, a bill of materials, and colorway options [Sesamers, November 2026]. This process is traditionally manual, iterative, and prone to error, often requiring multiple rounds of communication between design teams and manufacturers. By generating these packs from the AI-rendered design, Fermat is attempting to own a critical, non-creative bottleneck in the fashion supply chain.
Traction With Major Retailers
Fermat reports it now serves more than 350 customers across 14 countries [FashionNetwork, November 2026]. Its roster includes prominent names like ASOS, J.Crew Group, Puig, and New Look, suggesting the product has moved beyond early adopters to established enterprise buyers. The company claims its software cuts 75-80% off core design tasks and saves an average of over 100 hours per designer each month [Fermat, Unknown].
A look at the customer base reveals the product’s fit across different segments of the industry:
| Customer | Segment | Implied Use Case |
|---|---|---|
| ASOS | Fast Fashion / E-commerce | Rapid design iteration for high-volume cycles |
| J.Crew Group | Premium Apparel | Streamlining design-to-sample for branded collections |
| Puig | Luxury & Fragrance | Likely applied to apparel lines or brand extensions |
| New Look | Value Fashion | Cost and time reduction in product development |
Weekly active user counts are also substantial, with the company reporting more than 2,000 designers using the platform each week [Scommerce, Unknown]. As of March 2026, Fermat employed 83 people, indicating significant investment in product and go-to-market functions ahead of this fundraise [Revelio Labs, March 2026].
The $8 Million Expansion Bet
The recent $8 million seed round was led by Kibo Ventures and Lunar Ventures, with participation from PROfounders Capital, Masia Ventures, and several other European and angel investors [FashionNetwork, November 2026]. The capital has a clear geographic target: funding a new office in New York City. This move is a direct play for the American fashion and retail market, a logical step for a company whose existing enterprise customers already have major operations there.
The expansion is not just about sales. CEO and co-founder Pol Baladas has indicated the funding will support expanding Fermat’s product from a design-focused tool into a broader garment production operating system [FashionNetwork, November 2026]. This suggests ambitions to move further down the manufacturing pipeline, potentially integrating with sourcing, costing, or production management systems.
Technical Breakdown and Scale Risks
The technical premise rests on a pipeline of specialized models. First, image generation models trained on fashion-specific datasets create photorealistic renders. Second, and more critically, system must interpret those images to extract structured data,identifying seams, materials, and dimensions,to populate a technical spec sheet. The accuracy of this translation is everything; a mis-specified measurement can lead to an entire production run being scrapped.
The risks at scale are primarily about data fidelity and workflow rigidity.
- Specification accuracy. The AI’s interpretation of a visual design must be perfect. A ‘relaxed fit’ is subjective; the corresponding armhole measurement is not. The model’s tolerance for error is effectively zero for high-stakes production.
- Integration depth. The value multiplies if Techpacks integrate directly with factory PLM (Product Lifecycle Management) or ERP systems. Without those deep integrations, the output remains a PDF that must be manually re-entered elsewhere, blunting the efficiency gains.
- Creative guardrails. Designers may resist tools they perceive as overly prescriptive or limiting. The software must enhance creativity without imposing a rigid, system-defined structure on the early ideation phase.
Fermat’s answer to these risks appears to be its ‘Tool Editor,’ a no-code interface that lets design teams create custom AI models using their own moodboards and material libraries [Fermat, Unknown]. This offers a path to customization and control, allowing brands to tailor the system to their specific house style and quality standards.
The Next Twelve Months
The immediate milestone is the successful launch of the New York office, which will test Fermat’s ability to sell its European product into the competitive US retail technology landscape. Success there could trigger a larger Series A round within the next year to fuel further R&D and global sales expansion.
Investors are betting that the fashion industry, historically slow to digitize its core creative processes, is now ready for an AI-native operating layer. Fermat’s early traction with major brands suggests that bet might be right. The next phase will be determined by whether it can move from being a valuable design accelerator to becoming the indispensable system of record that connects the sketchpad to the sewing machine.
Sources
- [FashionNetwork, November 2026] Spanish company Fermat raises $8 million to fuel its expansion in the United States | https://us.fashionnetwork.com/news/Spanish-company-fermat-raises-8-million-to-fuel-its-expansion-in-the-united-states,1885706.html
- [Sesamers, November 2026] Fermat raises $8M seed for AI fashion software | https://www.sesamers.com/funding/fermat-raises-8m-seed-ai-fashion-sketch-to-factory/
- [Pomegra, November 2026] Fermat raises $8M to bridge the gap between fashion designers and manufacturers | https://tech.eu/2026/11/17/fermat-raises-8m-to-bridge-the-gap-between-fashion-designers-and-manufacturers
- [Scommerce, Unknown] Fermat Seed Funding: $8M for AI Fashion Design Start-up | https://www.scommerce.com/fermat-seed-funding/
- [Revelio Labs, March 2026] Headcount data for Fermat |
- [Fermat, Unknown] Company website and product claims | https://www.fermat.app/