Tech
Fashion startup Atorie raises $9.5M to bring consumers luxury goods without the markup
Fashion startup Atorie announced Thursday a $9.5 million seed round with investors, including a16z speedrun, Night Capital, and Lightspeed Ventures’ Jeremy Liew.
Shoppers can visit the Atorie website and buy handbags or clothes made from the same material — and coming from the same factory — that manufacturers use in high-end goods. The items are reasonably priced, too, with an Italian leather handbag costing just a few hundred, compared to the thousands a brand like Prada or Louis Vuitton would sell it for.
The startup arrives at a time when dupe culture has become increasingly popular, while the luxury sector has faced backlash from consumers in the post-pandemic era due to swift price hikes.
As a result, young consumers especially have sought cheaper, near-identical replications of these high-end goods; doing so has become almost a status symbol itself.
On Atorie, the items sold are mostly not dupes, says co-founder and serial entrepreneur Redouane Ramdani.
“It’s the same material, same craftsmanship,” he said. “It’s coming from the same factories.” He doesn’t consider Atorie fast fashion either. “It’s slow,” he clarified.
Before Atorie, Ramdani built the creator platform Snipfeed, which was acquired in 2024. Having grown up in France with a family that worked in luxury manufacturing, the founder always had an idea in the back of his mind that he would one day do something in the industry he grew up loving as a kid.
By the time he sold Snipfeed, however, the luxury manufacturing industry was quite different.
The biggest shift he noticed was that luxury factories were no longer just manufacturing goods. Traditionally, a brand like Ralph Lauren would bring its own designs and materials to a factory, which would then produce them. The problem was that brands had to commit to large minimum orders, which often pushed them to overproduce inventory. At the same time, a lot of these factories depend on working with a small number of large brand customers. If a brand pulled out at the last minute or not enough of that overproduced inventory sold, the factories faced financial and inventory risk.
“What’s changing is that the best factories increasingly have their own design and product-development capabilities,” Ramdani told TechCrunch. “Instead of simply manufacturing someone else’s designs, they can develop products themselves, adapt them quickly, and produce in smaller batches.”

AI also helps these factories by pulling data that helps them identify which products are likely to sell out before committing to large production runs.
“That reduces overproduction and allows factories to diversify beyond a handful of large customers,” he said.
These changes also gave Ramdani an idea, leading him to team up with Luis Angulo to launch the startup, which is now an AI-powered fashion brand that connects luxury manufacturers directly with consumers.
He compared his company’s approach to the retailer Quince, which is known for selling high-quality, low-priced items.
“We are talking to a different generation who is coming back for things that are trendy but well-made,” he continued, adding that consumers are getting tired of pure fast fashion, especially because of the harmful impact it’s having on the environment.
AI, of course, is used. Ramdani sees AI agents changing the entire shopping experience, where, in the future, people just instruct AI agents to buy things for them.
Ramdani said the company uses AI to analyze fashion trends and test what colors might look good on a product. It also uses AI agents to estimate consumer demand and trains the agents to predict when a material might start running short so factories can stock up in time.
On the consumer side, the Atorie website offers an AI agent that can build an outfit based on whatever — or whoever — inspires the customer. Over time, Atorie uses AI to learn shoppers’ habits to suggest what to buy next.
Ramdani said Atorie is already seeing an increase in sale referrals from platforms like ChatGPT and Claude.
The company said its goal is to become an alternative to Zara and offer something “very high quality for a price point that’s very affordable.” The company ended last year with around $5 million in sales and is expected to reach an annualized run rate north of $55 million this year.
“We are growing super fast,” Ramdani said, adding that the team is working with over 40 factories across the world right now.
The fresh capital will fund logistics, build more AI tools, and support production. The company also plans to release its own in-house line, like Amazon Essentials, and will work with creators and influencers to help them launch their own clothing lines quickly.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
>
Tech
Neocloud Lambda secures $1B in debt to buy more chips
Lambda, an AI cloud company that buys computing chips and rents them out to businesses, has raised $1 billion in private, short-dated debt to buy Nvidia’s AI chips that it will lease to Microsoft, Bloomberg reports.
The terms of the deal, which Bloomberg says was arranged by JP Morgan Chase, signal that Lambda is betting it will be able to quickly deploy the chips and start generating revenue from them, letting it repay the debt fairly quickly using that incoming cash.
This is the latest in a string of loans that Lambda is using to fund GPU infrastructure for specific customers. In May, it closed a $1 billion secured credit facility, and this week it announced the closing of a $926 million loan to fund Nvidia GB300 GPUs, one of Nvidia’s newest chip models, for a deployment it’s under contract to provide Nvidia itself.
The $1 billion private debt deal comes as as Lambda is reportedly in talks for a $3 billion pre-IPO round. The company last November raised $1.5 billion in venture capital at a $5.43 billion post-money valuation, per PitchBook data.
Lambda isn’t the only one relying on debt to fund the AI boom — according to data Bloomberg compiled, banks and tech companies have raised over $400 billion in AI-related debt globally in 2026 so far.
>
Tech
Oscar Winner Jernej Barbič Brings Monsters to Life
While walking to school as a child, Jernej Barbič would marvel at how beautiful his home was. He was born and raised in a picturesque village in northwestern Slovenia (formerly Yugoslavia), located in the European Alps. Surrounded by alpine and beech trees, Barbic dreamed of replicating their swaying in the wind for others to enjoy.
At the time, he didn’t have the tools or the knowledge to create a system that could do that, but it sparked his interest in computer graphics, he says.
Jernej Barbič
Employer
University of Southern California, in Los Angeles
Title
Professor of computer science
Member grade
Senior member
Alma maters
University of Ljubljana, in Slovenia; Carnegie Mellon
Twenty years later, in 2016, Barbič, a professor of computer science at the University of Southern California, in Los Angeles, made his mark. His Ziva VFX software system allows for the creation of realistic muscle, fat, and skin simulations for 3D digital humans and creatures.
The technology was launched in 2016 by a startup he helped found, Ziva Dynamics, headquartered in Vancouver. It was acquired in 2021 by Unity Technologies of San Francisco.
Ziva VFX has been used in more than 60 movies including Aquaman and the Lost Kingdom; Godzilla x Kong: The New Empire; and Venom: Let There Be Carnage.
For the design and development of Ziva VFX, Barbič, an IEEE senior member, received a 2025 technical achievement Academy Award. It was a “tremendous honor,” he says, as the award recognizes technologies that have had a significant impact on motion picture production.
“Computer graphics and simulation can sometimes feel like a specialized technical field,” he says, “but the award showed that these ideas affect not just science but also art and how stories are told on screen.
“The digital characters enabled by mathematics become important parts of people’s lives.”
Barbič says he was inspired to pursue engineering by his father, an engineer who headed a cement factory’s research department and invented a technology that uses magnetic resonance imaging to test the integrity of cement. His father’s work showed him that “mathematics and physics are beautiful on their own, but engineering lets you build something that other people can use,” he says.
It was Barbič’s mother, an elementary school teacher, who introduced him to computer science. When he was 8 years old, the school his mother taught at bought a ZX Spectrum computer. With permission from the principal, she brought it home for her son to play on for two weeks. But he didn’t just play games; he created his own game using the BASIC programming language.
The machine came with a booklet that contained instructions on how to write a computer program, he says.
“At first,” he says, “I copied them verbatim without understanding what they did. But then I started realizing there is structure, and I modified the instructions.”
Of all the creatures brought to life using his technology, Barbič is particularly enamored with King Kong from 2024’s Godzilla vs. Kong.DNEG/Warner Bros. Entertainment Inc./Legendary
By the end of the two weeks, he’d developed a computer game where players guided a snowman along a winding road. It shifted unpredictably to the left or right, and players accumulated points by remaining on the road for as long as possible.
Barbič went on to earn a bachelor’s degree in mathematics in 2000 from the University of Ljubljana, in Slovenia. The following year, he moved to the United States to begin a doctoral program in computer science at Carnegie Mellon. It was a major turning point in his life, he says.
His doctoral research focused on developing simulation methods for objects that can change their shape when an outside force is applied to them, known as “deformable objects.”
That project shaped much of his later research, he says: “I became interested not only in making simulations accurate but also in making them practical: fast enough, robust enough, and controllable enough to be used in real applications.”
After earning his Ph.D. in computer science in 2007, he worked as a postdoctoral researcher at MIT. Two years later, he joined USC as an assistant professor.
Making movie magic possible
It was at USC that Barbič merged his passion for computer science with film. He developed Vega FEM, an open-source software program that allowed people to animate realistic 3D deformable objects. But Vega FEM was narrow in scope and not exactly what filmmakers needed, he says, so he started exploring how to create a version suitable for the movie industry.
“A major theme of my career has been the translation of research ideas into practical tools,” he says. “Academic research often produces beautiful algorithms, but it can be difficult to make those algorithms usable by artists, engineers, or production teams. I have always been interested in that bridge: taking rigorous computational methods and turning them into systems that people can actually use.”
In 2011 he attended an Association for Computing Machinery conference presented by its Special Interest Group on Computer Graphics and Interactive Techniques (SIGGraph). There he met James Jacobs, the creature supervisor at visual effects company Weta FX of Wellington, New Zealand. The company is behind the effects in the Lord of the Rings and Hobbit trilogies and other movies. Jacobs used Barbič’s software to create animals and fantastical creatures.
Two years later, Weta FX offered Barbič a summerlong research position in New Zealand. He accepted and spent the time studying the process of creating visual effects and learning what roadblocks existed in the film industry, he says.
At the time, the technology to create realistic soft-tissue and anatomical simulation for digital characters didn’t exist.
“The visual effects industry had reached a point where surface-level realism was not enough,” Barbič says. “A creature could have beautiful skin textures and detailed geometry, but if the bones, muscles, and fat underneath did not move correctly, the illusion would break.
“The problem was especially difficult for creatures and characters that need to feel alive: animals, monsters, fantasy creatures, or digital doubles. Their bodies may have unfamiliar anatomy, but the audience still anticipates them to move in a way that matches real-world expectations. Muscles should bulge and contract, skin should stretch and slide, fat should have inertia, and tissue should respond to motion and impact.”
In an effort to solve the problem Jacobs in 2014 approached Barbič about founding a startup. In 2015 they launched Ziva Dynamics, where they began what is now Ziva VFX.
The software uses physics-based simulations to model the internal anatomy of a character. It numerically solves the partial differential equations of nonlinear elasticity for musculoskeletal human and creature tissues, Barbič says. The equations describe how muscles, fat, skin, and connective tissue deform, interact with bones, and connect, and how muscles activate. Instead of animating only the outside surface, artists can create a model with underlying muscles, bones, soft tissue, and fat. Each component is assigned material properties, constraints, attachments, and activations. The simulator then computes how they deform and interact over time.
The technology uses ideas from computational mechanics, finite element methods, numerical optimization, contact handling, and computer graphics, Barbič says. Finite element simulation, a method used to predict how a product or structure reacts to heat and other real-world forces, provides a way to model deformable materials volumetrically, not just as surfaces, he says. The tool computes internal elastic forces and solves the equations of motion so the character’s tissues respond plausibly to animation, pose changes, muscle activation, and dynamic motion.
But the system had to be designed for artists, Barbič says. In production, he says, the goal is not only physical realism but also controllable realism.
“Artists need to direct the result, iterate, and fit the simulation into a larger animation pipeline,” he says. “So the technology had to combine scientific simulation with practical controls, robustness, and integration with visual effects workflows.”
Barbič says Ziva VFX has been used in more than 60 movies. Of all the creatures brought to life using his technology, he is particularly enamored with King Kong from 2024’s Godzilla vs. Kong.
“When King Kong is walking, you can see the muscles, how they’re very pronounced, and how they influence the shape of the skin. You can really feel the strength of King Kong,” he says. “And this was made through my software, so I think it’s amazing.”
After Ziva Dynamics was acquired by Unity, Barbič consulted for the company for almost two years.
In 2024 DNEG, a London-based visual effects and computer animation company, acquired the exclusive license to Ziva VFX.
Animating the human hand
Barbič strives to improve visual effects as an entrepreneur and an academic. His most recent research, funded by the U.S. National Science Foundation, focused on the modeling, simulation, and animation of human hands. The goal is to create computer models of hands that can be used to design tools, medical prosthetics, and robotic hands.
“The hand is a fascinating and difficult system,” Barbič says. “It contains many small bones, muscles, tendons, ligaments, skin, fat, and other soft tissues, all packed into a compact structure and interacting mechanically in complex ways.”
He and his team built a digital twin of the human hand.
He aimed to move toward “anatomically meaningful simulation,” he says. He used medical imaging, geometric modeling, finite element methods, and multibody simulation to represent the internal structures of the hand and its motions.
“IEEE lets me place my work not only in the world of images and animation but also in the world of engineering systems that must be accurate, stable, interactive, and useful.”
He worked with Bohan Wang, who at the time was a USC doctoral candidate, and George Matcuk, an associate professor of radiology. Wang is now an assistant professor of computer science at the National University of Singapore.
Barbič, Wang, and Matcuk scanned four people’s hands with an MRI machine. The two men and two women would position their hands in 12 poses, which allowed the team to gather data about how the bones, muscles, and fat move with each pose.
The data sets are available for anyone to use in their own studies.
“This project can help medical doctors learn more about how the hand is moving,” Barbič says. “It’s also great for roboticists to better understand how the human hand actually works, so [the movements] can be replicated.”
IEEE: Integral in interdisciplinary research
Barbič joined IEEE in 2008, when he published his research paper on simulation methods for deformable objects in the inaugural issue of the IEEE Transactions on Haptics. He has since published several papers in the IEEE Transactions on Visualization and Computer Graphics, which he says connected his work to a wider community interested in visual computing and computational methods. You can find his research in the IEEE Xplore Digital Library.
“IEEE recognizes the engineering side of computer science,” he says. “My work is often presented as computer graphics, but at its core, it is also simulation, mechanics, numerical methods, haptics, visualization, and software systems.
“IEEE is a community where that broader identity makes sense. It lets me place my work not only in the world of images and animation but also in the world of engineering systems that must be accurate, stable, interactive, and useful.”
He believes the organization is key in supporting a healthy interdisciplinary research ecosystem at a global scale—which, he says, is why he has served as an associate editor for Transactions on Visualization and Computer Graphics and Transactions on Haptics.
Being a member has made it easier for Barbič to connect with engineers in different fields, he says.
“My research often lives between categories: It is mathematical but also practical; visual but also mechanical; artistic but also engineering-driven,” he says. “IEEE is one of the professional communities where that mixture is understood.”
From Your Site Articles
Related Articles Around the Web
>
Tech
An Anthropic researcher just gave us a peek at self-improving AI
Training AI models with other AI models has become a very popular goal for neolabs — and now, a researcher in Anthropic’s fellows program has given us an early look at what it might look like in practice.
On Friday, Anthropic published a new paper titled “Automated Researchers Can Reliably Mitigate Alignment Failures,” detailing how AI systems could reliably improve a model’s performance on a set of alignment benchmarks. When given 10 benchmarks for specific misaligned behaviors, the automated systems were able to improve performance on every single one without degrading overall performance.
Led by Anthropic Fellow Chen Yueh-Han, the system replicates much of the traditional approach to research. Each automated system searches the available literature, proposes a method, and trains the model using that method for 30 minutes, gradually increasing the benchmark over several iterations. Effective methods are preserved while ineffective ones are discarded, allowing the system to operate quickly and at a great scale.
“Overall, these results provide early evidence that automated alignment post-training could become practical in the near term,” the paper reads.
The paper is a step toward recursive self-improvement, which many see as the next significant step in AI progress. If models can improve their own alignment training, it’s plausible they could improve training practices more broadly — at which point, human AI researchers might soon become obsolete.
The paper isn’t shy about addressing this idea, explicitly comparing the Automated Alignment Researcher (AAR) to its human equivalent. “The best AAR method beats what experienced humans propose, on average within six hours,” the paper reads. “Human guided research directions do not lead to stronger performance.”
There’s even a cost comparison, in case anyone wasn’t convinced. “An AAR costs roughly $4 per hour in API inference against the $150 per hour we pay our human researchers.”
In fairness, the paper also points out a few limitations to this approach. The automated system only works insofar as the benchmarks reflect the actual alignment goals, and even then there’s significant work to be done in establishing and maintaining those benchmarks — not to mention maintaining and expanding on the literature the automated researchers are draw from.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
>
-
movies3 months agoSearch For Canadian TV Actor Stewart McLean Now Homicide Investigation
-
Fashion9 years agoThese ’90s fashion trends are making a comeback in 2017
-
Fashion9 years agoAccording to Dior Couture, this taboo fashion accessory is back
-
Fashion9 years agoModel Jocelyn Chew’s Instagram is the best vacation you’ve ever had
-
Fashion9 years agoEmily Ratajkowski channels back-to-school style
-
Fashion9 years ago9 Celebrities who have spoken out about being photoshopped
-
Fashion9 years agoYour comprehensive guide to this fall’s biggest trends
-
Anime3 months agoRurouni Kenshin: Hokkaido Arc Manga Takes 1-Issue Break – News
