reverse.fashion adds HTGF backing to automate textile sorting

Berlin-based reverse.fashion, a TU Berlin spin-off using AI for used-textile sorting, secured a seven-figure pre-seed extension from High-Tech Gruenderfonds. The funding will support commercial expansion of its computer vision and deep learning systems as EU textile regulations drive demand for automated sorting solutions.

Mario Osterwalder, Dr. Karsten Pufahl and Paul Doertenbach have secured a seven-figure pre-seed extension from High-Tech Gruenderfonds https://www.htgf.de/en/venture-capital-investor-2/?ref=runtimewire for reverse.fashion https://reverse.fashion/?ref=runtimewire , the TU Berlin spin-off building AI systems for used-textile sorting, Tech.eu reported on July 8th, 2026 https://tech.eu/2026/07/08/berlins-reversefashion-bags-seven-figure-funding-to-scale-textile-sorting/?ref=runtimewire . The amount is deliberately imprecise. reverse.fashion and HTGF disclosed only that the extension is in the seven figures. They did not disclose valuation, the exact close date, ownership, revenue, headcount, customer count or total capital raised to date. That matters because the funding story here is less about the size of the round than about what investors are underwriting: a bet that textile sorting, one of the least glamorous parts of circular fashion, is becoming a software-and-automation problem. reverse.fashion was founded in 2024 by a team that maps unusually well to that problem. Pufahl is the technical founder tied to years of textile-sorting research. Doertenbach brings sorting-floor operating experience, including more than a decade running a modern garment sorting facility and work on European textile extended producer responsibility. Osterwalder previously co-founded circular.fashion and has worked across circular economy, textile sorting and Digital Product Passport projects. The result is a company built around a narrow insight: the value of a used garment is often decided in seconds, by a sorter making a judgment call under volume pressure. reverse.fashion wants to replace that judgment call with an industrial data layer. Its system uses computer vision, deep learning, Digital Product Passport integration and sensing technologies to classify used garments by condition, defects, brand, style, size, material composition and routing option. In practice, reverse.fashion is selling textile operators a way to decide whether a shirt should go to resale, repair, upcycling or recycling, with the decision recorded and repeatable rather than trapped in a worker's eye. The new money follows a strategic pre-seed The HTGF extension comes after reverse.fashion completed a six-figure pre-seed financing in October 2025, led by KISORA with H&M Group among the other investors, according to legal adviser Heuking https://www.heuking.de/en/news-events/latest-news/article/heuking-advises-berlin-based-fashion-logtech-start-up-reversefashion-on-its-pre-seed-financing-round.html?ref=runtimewire . That earlier investor mix already pointed to the commercial pull: KISORA sits close to textile logistics, while H&M Group has obvious exposure to the cost and reputational pressure around apparel waste. HTGF's check adds a different kind of validation. HTGF https://www.htgf.de/en/venture-capital-investor-2/?ref=runtimewire positions itself as a pre-seed and seed investor for German technology startups. For reverse.fashion, the extension is intended to fund commercial expansion, including continued deployment of co.sort software and rollout of the automated line.sort system. The timing is not accidental. The European Union's targeted 2025 revision of the Waste Framework Directive entered into force on October 16th, 2025, with textile extended producer responsibility at its center. The European Commission says the revision requires member states to set up EPR schemes under which textile producers contribute to management of used and waste textiles, and it aims to harmonize the market for used and waste textiles across the EU. That creates a direct incentive for brands, collectors and sorters to know what is in the waste stream, where it goes and what value can still be recovered. Digital Product Passports are part of the same pressure. A T-REX Project webinar summary https://trexproject.eu/news-article/the-importance-of-data-and-digital-led-solutions-for-scaling-circularity-in-the-fashion-industry-webinar/?ref=runtimewire framed DPPs as a tool for transparency and traceability that will be required by multiple incoming EU regulations. reverse.fashion's DPP integration matters because sorting decisions become more valuable when they can combine what a garment looks like today with verified product data from its first life. What reverse.fashion says it has built reverse.fashion has two named products. line.sort https://reverse.fashion/?ref=runtimewire is the automated system: it captures garments on a conveyor, makes sorting decisions in one pass and routes pieces into bins. co.sort https://reverse.fashion/?ref=runtimewire is the human-assist product: it recommends styles, fractions and sorting outcomes while the human sorter makes the final call. Both run on reverse.engine, reverse.fashion's AI and software stack. On its website, reverse.fashion says its models are trained on millions of textile images, can grade garment condition against 20-plus defect types, detect defect size and location, recognize trims, prints and brand labels, support unlimited sorting categories and run at conveyor speeds of up to 2 meters per second. reverse.fashion also claims 40% higher productivity versus manual multi-step sorting and 20% projected cost reduction. Those numbers should be read as company and pilot claims rather than audited operating results across a broad customer base. reverse.fashion names Deutsche Kleiderstiftung as a customer since 2025 and quotes its managing director, Ulrich Mueller, emphasizing that declining textile quality makes human-machine collaboration necessary. reverse.fashion has not published a customer list, retention metrics, pricing or revenue. A May 2026 Fashion for Good and Circle Economy Project Rewear report https://ce-assets.prod.circularity-gap.world/FFG x Circle Economy Project Rewear fa69e780ee.pdf?ref=runtimewire included a reverse.fashion case study that described line.sort capturing near-360-degree images and classifying items in real time at conveyor speeds up to 2 meters per second. The report highlighted productivity gains and revenue uplift in pilot settings, while noting that findings were directional rather than definitive and that brand-specific recovery remains to be tested. The same report flagged the social cost of automation. Sorting roles are often held by people distant from the labor market, and automation can reduce short-term personnel needs. That does not undercut reverse.fashion's pitch. It defines the implementation problem: if the software works, operators still have to retrain, redeploy and manage the transition rather than treating labor savings as the whole business case. The market is moving from fiber detection to value detection reverse.fashion is entering a field where automated textile identification already exists, but much of the established technology has focused on fiber composition and recycling feedstock. Valvan's Fibersort https://www.valvan.com/en/solutions/textile-sorting-recycling?ref=runtimewire uses AI models, near-infrared spectroscopy scans and RGB camera data to separate post-consumer textiles by fiber composition and color. Refiberd https://refiberd.com/technology/?ref=runtimewire uses AI-based hyperspectral imaging to predict fiber composition and detect contaminants in textile waste. reverse.fashion is positioning itself closer to the reuse and recommerce decision. Material composition still matters, but resale value can depend on smaller and messier signals: whether a garment is a premium brand, whether defects are visible, whether a trend is moving, whether a sorter can bundle a coherent fraction for a buyer, and whether the product data can support resale or compliance. That is a harder classification task because it blends physical detection with market context. That is also where Osterwalder, Pufahl and Doertenbach's founder mix matters. A pure AI team could underestimate how chaotic used-textile flows are. A pure sorting-operations team could underestimate how quickly regulation and recommerce data requirements are changing. reverse.fashion is trying to join the sorting line, the compliance layer and the resale market in the same product. The open question is how quickly that product can move from promising deployments to repeatable industrial sales. reverse.fashion has one named customer on its site and a fresh HTGF extension, but no disclosed revenue base. The capital should buy time for line.sort deployments and co.sort software adoption. It will also test whether European sorters will pay for automation before EPR fees and DPP obligations fully settle into national implementation. If reverse.fashion is right, the sorting facility becomes a data facility. Every garment that passes through line.sort is not only routed to a bin. It becomes a classified asset with condition, quality, brand, defect and destination data attached. That is the layer brands, recyclers, recommerce platforms and regulators will need if circular fashion is going to operate at industrial volume rather than pilot-project scale.