Here are the fashion problems AI can solve, according to investors
By Maghan McDowell, Vogue Business
Funding for startups that use artificial intelligence — and particularly generative AI — is booming.
In 2023, more than 400 AI companies raised funding, totalling more than $21 billion globally, says CB Insights; in 2022, funding for AI totalled $4.3 billion across 257 total deals. Momentum has continued this year, with many of the most well-funded companies offering a range of generalist tools. OpenAI, which owns text-to-image and text-generation tools ChatGPT and Dall-e, raised at least $13 million from Microsoft alone, and recently raised another undisclosed amount that values the company at $80 billion. Stability, an open-source AI company that offers a range of generative AI tools, just added another undisclosed amount to its previous $101 million fundraise.
The next step? Investors are looking at individual industries to find, and fund, the AI tools that will help solve specific problems. Companies that apply generative AI solutions built specifically with fashion in mind are hoping to ride the wave of investor enthusiasm. “Our focus has been identifying solutions that amplify design capabilities and maximise efficiency gains at every step of the design and sales process,” says Dana Settle, co-founder and managing partner at venture capital firm Greycroft, which just invested in fashion AI company Raspberry.
Already, fashion is one of the first industries to experiment with generating both text and imagery, ranging from the fashion design process (Revolve), advertising content (Prada Beauty and Chanel Beauty), customer service (Zalando) and SEO-friendly text for product descriptions (Adore Me). Gucci is commissioning artists to create art using generative AI that uses Gucci’s intellectual property, and parent company Kering is experimenting with a chat tool that can make product recommendations. LVMH is working with Stanford University’s Institute for Human-Centred Artificial Intelligence to explore AI applications.
Brands and retailers are keen for more solutions that use AI, says Jackie Trebilcock, managing director of retail accelerator New York Fashion Tech Lab, which matches startups with established brands and retailers who pay to participate (past participants include LVMH, Richemont and Burberry). AI was among the top themes for this year’s applicants, she says. While the cohort is not yet announced, the selected companies will cover areas including AI for design, merchandising, personalisation and visual commerce, Trebilcock says.
“Now that AI has become a saturated space, we look for founders building nuanced, verticalised applications of AI in addition to exceptional founder-market fit,” says Bailie Salk, junior partner at Outlander VC, which has been investing in early-stage AI startups for most of the last decade, Salk says. These include Scale AI and Imbue, which are now “unicorns” valued at more than $1 billion.
A new batch of companies receiving early funding suggests that investors see an opportunity to further democratise access to design and improve the speed and quality of creation, while reducing waste. They also see an opportunity for customer-facing tools that help people find products or acquire personalised products. While it’s still early days — ChatGPT only became available in late 2022 — already, a number of startups have begun emerging out of stealth mode and have attracted big investors and pilot partnerships with brands, offering a first look at the problems and opportunities that investors are eyeing in fashion’s AI renaissance. Here are six to know.
Raspberry AI: Creating trend-driven designs
Before founding Raspberry AI in late 2022, Cheryl Liu’s work as an analyst at global investment firm KKR saw her frequently evaluating and learning about various fashion business models. When generative AI companies hit the mainstream, she saw an opportunity to apply the tech more specifically to fashion. Raspberry AI uses generative AI to help design fashion, using a demand trends platform and a multimodal design tool that enables users to translate trends into clothing prototypes.
The company has raised $4.5 million in seed funding from a list of investors that includes Khosla Ventures (which also invested in ChatGPT and Dall-e parent company OpenAI), Greycroft Partners, Revolve co-founder and co-CEO Michael Mente and Reformation founder Yael Aflalo, among others. It recruited Fan Yang, who trained GenAI models for Google, to be head of AI, and now counts H&M, Zara owner Inditex as a client, in addition to smaller brands.
Raspberry frees up time for creativity and addresses key industry challenges, while helping increase sell-through, revenue and sustainability, says Greycroft’s Settle. Raspberry also helps “bridge the gap between hobbyist-level art generation tools and enterprise solutions for designers. Its AI model understands fashion-specific terminology,” Settle adds.
“Our investors were very bullish on vertical applications of Gen AI,” including fashion, music, art, architecture and marketing, says Liu. The thesis of the investment community now, she says, is to take generalised large language models and apply them to specific problems and industries. Otherwise, basic tools can require significant training to be useful. “We felt that fashion was one of the most under-hyped applications to apply the tech to, because there are so many new designs that are demanded from the industry; it seemed like a perfect use-case for AI,” she says, referring to the fast-paced trends cycle and the increase in styles and seasons.
Liu describes Raspberry as somewhat like WGSN-meets-Midjourney, referring to the consumer trends platform and the image-generation software — but optimised for the fashion industry. While there are many text-to-image generation tools, Raspberry is specifically trained to better understand the needs of designers and brands, which streamlines the process of using generative AI in fashion.
A key goal was to help brands keep up with the trend cycle. “One of the biggest problems over time was that more and more inventory was left unsold,” Liu says. “There was a bigger disconnect between supply and demand from the customer, largely driven by a shift to social media and people getting inspiration and ideas from social and Google than in-store.” She adds that while it can speed up the design process, the tool is designed for any fashion brand — not just fast fashion.
Flock AI: On-model photography
Flock (pictured at top) enables brands to generate and edit on-model photography that is in keeping with a brand’s aesthetic guidelines; a brand can input content such as a flat-lay garment, tech pack or digital pattern (or makeup colour) to generate an on-model image, and test how details such as hairstyle, pose or background are impacting shopper behaviour. The company plans to announce a pre-seed round of $1.5 million this month.
Flock's founders were core to the investment, says Outlander VC’s Salk. Outlander VC looks for early-stage founders with “extraordinary vision, intelligence, character and execution”, Salk says. “Across all four categories, Vita and Malavika score exceptionally high, representing the perfect blend of industry expertise and technical acumen.” Salk adds that Flock’s technology enables companies to diversify from “one-size-fits-all imagery”. “While truly inclusive photoshoots remain impossible at scale, Flock is tackling the demands for diversity and inclusion with its unique applications of generative AI.”
Co-founder and CEO Manvitha Mallela, a former buyer at Bloomingdale’s, says that Flock’s AI tools are a way to improve what she found to be an inefficient process of creating product imagery for the website. She and co-founder Malavika Reddy, whose background is in computer vision, are also both women of colour who “have always struggled to see brand content that related to us and made us confident in making purchasing decisions online”, Mallela says.
Blng.Ai: Democratising jewellery design
Blng.Ai was founded in June 2023 by Valerie Leblond to apply generative AI to jewellery design. It has recently raised seed funding of at least $2 million from a strategic industry partner and angel investors. The company was chosen by LVMH to participate in a startup showcase, and has attracted clients including a celebrity brand, a major retailer and a large diamond house as clients.
The idea of Blng was sparked when its CTO (now Leblond’s husband) began the slow process of designing a custom ring with a luxury brand. The resulting multimodal software that Leblond’s team created — many are former Disney engineers — can convert rough sketches and text prompts into photorealistic three-dimensional renders of jewellery — making a process that typically takes three weeks almost instantaneous.
“Their technology empowers creativity at the speed of imagination. This means technical skills like 3D modelling and rendering are no longer a bottleneck in the jewellery design process,” says Blng investor Tara Tan, who is a general partner at Strange Ventures. Tan adds that while there are many generative AI startups, Blng stands out because of its commitment to craftsmanship and quality.Now, the technology can convert anything from an artistic gouache painting by a Bulgari artist to a simple sketch by an untrained person into lifelike visuals. Designers can then tweak the designs with text prompts, such as changing the material from platinum to gold, or making certain stones bigger. Once the design is ready, Blng also enables users to generate on-model photography and videos of the piece. “Its practical use of AI solves real-world problems of rapid-prototyping, precision marketing and custom manufacturing,” says investor David G. Wilson, GP of Seam Ventures.
Going forward, Blng will be building out its customer-facing shopping app, working to develop the capability for the tech to also understand the physics of physical pieces and look how it can extend towards on-demand manufacturing.
Mmerch: One-of-one, Web3-linked fashion, co-designed by artists
Mmerch, founded in 2022, is centred around the concept of “neo-couture”, which refers to pieces that are individual to the wearer but more accessible, both from a financial perspective and a manufacturing one, as individual designs are generated using generative design and modularised pieces. They are also connected to an NFT via an NFC chip in each garment.
This month, Mmerch announced that it had raised $6.4 million in seed funding, led by Liberty City Ventures, with participation from 6529 Holdings LLC, Christie’s Ventures and Flamingo DAO. Other participants include Karlie Kloss, Tory Burch and Web3 fashion consortium Red DAO.
Christie’s Ventures head Devang Thakkar says that founder Colby Mugrabi, whose background is in fashion and art, was as much of a draw as the business idea, especially in such an early stage. “It’s cliche in this early stage that the investment we make is in the founder, but the seed stage is so early that these concepts will evolve or pivot 10 times before it becomes a thing. Colby has a deep passion for this space, and the commitment and the hustle is quite strong in Colby.” He also liked that Mmerch enables artists to expand into fashion with generative art, and the potential in the near future for customers to wear digital versions of Mmerch products via mixed reality.
While the first collection serves as a proof of concept, Mugrabi plans to expand operations and bring in artist and brand partners on future additional drops, where artists can generate versions of their work to create future one-of-one collections.
Glaze: AI prompts to find shoppable products
Glaze enables people to use conversational text prompts or inspirational imagery to find something to buy online, merging the concept of a social media search with a ChatGPT prompt.
After graduating from Y Combinator in 2023, the company raised an undisclosed amount from investors who were interested in the future of shopping for Gen Z, says co-founder Maggie Zhang (all three co-founders are Gen Z women). The company is still in beta testing with a couple thousand users and partnering with brands, and plans to release the platform more widely in the coming months.
The problem of discovery online has become especially acute in the TikTok era, Zhang says.
“It’s more important for young shoppers to do a lot of social research to source inspiration and reference what is trendy. Search and inspiration no longer begins on a Google query or at the retailer’s website, but instead through content and at the recommendations of other people. Investors are really keen to see how AI can be leveraged as a tool to create a whole new model of shopping that is uniquely catered toward this generation that is still authentic to their self-expression.”
Kopia: Virtual clothing try-on
Kopia is the latest in a spate of companies that are using AI to enable customers to virtually try on clothing, and was recently selected to participate in Y Combinator, an elite startup accelerator that has graduated companies including Airbnb, Reddit and Dropbox.
Kopia, like all companies accepted to YC, is in the very earliest startup stages. Co-founders Aadi Nashikkar and Calvin Chen are also well positioned to apply genAI to try-on: Chen previously worked on e-commerce software and luxury apparel rental startups, and Nashikkar previously built multiple products using gen AI. Co-founder and CEO Calvin Chen says that using diffusion models — the type of generative AI that is able to create highly realistic and detailed imagery — is an upgrade because of its realism and ease of use.
To use Kopia’s technology on e-commerce sites, customers answer a few questions about their body type, input a selfie or headshot, and then can visualise themselves across any brand that uses Kopia’s technology and get size recommendations. It publicly launched earlier this month and already has interest from brands ranging from Gucci and H&M to Taylormade, Chen says. He says that accelerators like Y Combinator are looking for how AI can change massive industries, including fashion and e-commerce.
The next phase for the company will be continuing to fundraise, and showing off how VTO has come a long way, Chen says. “People have been sceptical of the idea of virtual try-on, as it’s been tried many times before, but we believe with AI, it’s actually possible and it will shift the entire e-commerce market.”