a day ago
Byte-Sized AI: Perfect Corp. and Nvidia Team Up; LuminX Gets Seed Round
Byte-Sized AI is a bi-weekly column that covers all things artificial intelligence—from startup funding, to newly inked partnerships, to just-launched, AI-powered capabilities from major retailers, software providers and supply chain players.
AI platform LuminX announced early this month it had closed a $5.5 million seed round, supported by 1Sharpe, GTMFund, 9Yards, Chingona Ventures and the Bond Fund.
More from Sourcing Journal
Chain Reaction: Dispatch Science CEO Arthur Axelrad on Turning Logistics into a 'Customer Experience Engine'
Inside VivaTech: LVMH Spotlights AI and Sustainability at Innovation Awards
Study Finds Machine Learning Can Cut Fabric Dyeing Waste
The San Francisco-based startup aims to create efficiency for logistics, supply chain and warehouse management functions. LuminX uses a combination of hardware and AI systems to allow the cameras it deploys to understand what's occurring in a facility in real time. The devices can be used at docks, on forklifts and more.
The devices have the ability to identify products and pallets, track their movement through the warehouse, understand the condition of packages and more. From there, the devices' findings are processed through an AI system that communicates with a human operator about the state of the warehouse. LuminX contends that the technology can decrease the amount of manual work necessary and provide direct recommendations for reducing errors and inefficiencies.
The startup plans to use the funding to add roles to its engineering team, scale up its go-to-market operations and refine the development of its technology for edge deployment.
Alex Kaveh Senemar, the company's CEO and founder said the seed round will help expand LuminX's footprint in the supply chain and logistics realms.
'This pivotal funding allows us to scale our next-generation AI models, transforming how warehouses operate,' he said in a statement. 'Our edge-based vision language models represent a massive step forward, acting as an intelligent core for warehouse operations. They deliver new levels of automation and insight, helping to turn previously opaque processes into transparent highly efficient systems.'
Walmart announced last summer that it was developing and beta testing a consumer-facing shopping assistant, designed to better guide consumers' e-commerce journeys.
The retail giant announced last week that the assistant, which it calls Sparky, is now publicly available to consumers using its app. Sparky seems akin to Amazon's version of a shopping assistant, which it calls Rufus. The e-commerce behemoth made Rufus publicly available in September 2024, and the two companies continue to compete over technology—both that directly aiding consumers and the systems optimizing their operations behind the scenes.
Sparky capitalizes on several of Walmart's existing AI capabilities, including review summaries. If a consumer asks Sparky whether a sundress will help keep them cool, the assistant can query existing reviews to give the consumer an idea of what other consumers have previously said about the dress' breathability and comfortability.
Walmart also previously enabled AI-assisted search, which Sparky further takes advantage of; instead of starting their shopping journey in the app's search bar, consumers can ask Sparky for product recommendations. For instance, a consumer might ask, 'Could you please help me find the best back-to-school styles for my third grader?'
In its announcement, Walmart contends that it intends to enable Sparky with agentic capabilities at some point.
'Ultimately, Sparky will solve everyday problems, freeing up time for what matters most. 'What's for dinner?' becomes a week of family-approved meal plans with ingredients automatically added to cart,' the company said. ''I need party ideas!' sparks a complete event planning session—theme, decorations, food and gifts—all highly personalized and coordinated within budget.'
Those types of queries—'What's for dinner? Add the ingredients to my cart,' for instance—when posed to Sparky today, prompt a message telling the user that the assistant isn't able to help with the request. While discussion around planning for agentic AI is on the rise, many companies lack execution on agentic to date.
Perfect Corp. announced this week that it has inked a computing partnership with technology giant Nvidia. Perfect Corp. offers virtual try-on, AI-powered recommendations, beauty simulation tools and other related applications to companies looking to make the e-commerce journey easier for their customers.
But today, one of the most difficult pieces of generating real-time imagery like this is that it doesn't always fully reflect how a product will appear in real life. The new partnership with Nvidia is meant to help bridge those gaps, helping Perfect to achieve better color representation, more accurate renderings, faster performance and more.
The company will leverage Nvidia TensorRT to speed up the rendering process, making the results feel near instant to consumers. It will also use Nvidia GPU capabilities, which help the models better create inferences about what a product will look like on an individual, helping to avoid the paper doll-esque virtual try-on experiences that so many consumers face in their e-commerce journeys.
Alice Chang, CEO and founder of Perfect Corp., said she expects the partnership to accelerate the company's impact on the fashion, retail and beauty industries.
'We pioneer sustainable, high-touch, AI and AR-powered omnichannel beauty and fashion tech solutions that empower brands to enrich their relationships with customers,' Chang said in a statement. 'Our collaboration with NVIDIA is a pivotal step in pushing the boundaries of what's possible in beauty and fashion tech, allowing us to deliver even more immersive and personalized experiences to consumers worldwide.'
PhotoRoom, which uses generative AI to aid sellers and brands in creating product images—in particular for marketplaces, like Depop—announced late last month it had acquired GenerateBanners. The financial details of the transaction were not disclosed.
The acquisition allowed the European startup to launch a new feature, called Visual Ads Automation. The tool allows PhotoRoom clients to easily generate advertisements for social, digital, email and more, customizing the content for individual consumer brackets or international markets. PhotoRoom allows customers to remove existing backgrounds and generate new backgrounds, add shadows and other details to match the new background and more.
Now that it owns GenerateBanners, it has enabled clients to use those images to generate ads directly inside the platforms. Clients can load their own colors, fonts, logos and more to help the systems adhere to existing brand style, and the new tool can create batches of advertisements simultaneously.
Matt Rouif, Photoroom's co-founder and CEO, said the platform's newly announced capabilities sweetens the deal for brands looking to reach customers and new markets in a scalable way.
'When you need the same product shot in ten languages and fifty formats, traditional creative workflows break.' Rouif said in a statement. 'Photoroom is becoming the AI creative agency for millions of businesses. Our new API lets teams generate on-brand visuals programmatically, at any scale, slashing costs and boosting performance.'
The announcement of the acquisition comes several months after PhotoRoom announced a $43 million Series B round.
