They say online media is going to be 3D, starting with e-tailing. Where can a merchant get that multi-dimensional content? They have 2D pictures, and they have physical objects. Occasionally there is a CAD. Creating 3D objects of the products is a hurdle companies must overcome, but their options are not great. Scanning? Expensive. CAD conversions take hours to be “filled in.” Converting 2D photos can be also done by remote contractors. All too expensive. Shopify is now AR enabled, and anyone can now add 3D inventory, or pay Shopify $300 to turn 2D images into 3D models. There is no good solution. A new start-up emerged from stealth this week, Threedy.ai (formerly AARWILD, Inc.). By using AI and computational geometry Threedy.ai can generate high-quality 3D models from 2D photos at a third of the cost and a fraction of the time current solutions require.
The expense and difficulty of creating 3D models of SKUs at scale discourages retailers from using 3D shopping to increase engagement and sales. The process is usually outsourced to countries like the Philippines and Vietnam. The cost has gone up considerably in the past five years as the growing demands of the spatial web has increased demand. To a lesser extent, manufacturers also do 3D scanning. William Sonoma recently acquired merchandise scanning startup Outward for $110M. Scanning results in higher quality, but it is not faster or less expensive than graphics solutions. Also down-sampling a high-poly model from the scanner to a low-poly model suitable for XR is not trivial. The other problem is that many e-commerce sites simply don't have the physical object at hand.
Retailers know that 3D shopping experiences are powerful tools to increase consumer engagement and online sales—particularly with home goods, furniture, and decor. AR-enabled shopping apps offered by Houzz, Ikea, Wayfair, and Overstock all allow consumers to preview products in 3D with AR before purchasing. Just months after launching their AR mobile shopping app, Houzz reported that app users who viewed a product in AR were 11x more likely to make a purchase than those who did not.
“Our vision is to become the Getty Images of 3D models for commercial products,” said founder and CEO Nima Sarshar, former Apple engineer who worked on AI. “Imagine a 3D model for every item on Amazon! We would like the enterprises to focus on the experiences they would like to provide, AR shopping, VR stores, and virtual showrooms, and not to worry about where they can get the 3D models required.”
The company’s vision is not limited to the products that line retail shelves. Sarshar believes everything that is 2D will soon become 3D. The opportunity, he says, is enormous. “On the spatial web,” he points out, “Zillow would be 3D.” The company is also interested in creating a digital asset management platform for its clients.
“I talked with most major retailers, including Walmart, Houzz, and Target. Everyone is excited about the potential of AR shopping, but it's clear that 3D content is the bottleneck. I founded Threedy.ai last August and left Apple to solve that problem using AI. The 2D to 3D conversion is our first product, but we are planning to fundamentally solve the 3D content problem. We have products for CAD-authoring and smart 3D scanning in the pipeline."
In addition to founder and CEO Sarshar (formerly part of Apple News’ Machine Learning group), the team consists of Mehdi Mohseni (Ph.D., Stanford) and Reza Jamei (Ph.D., Stanford, ex-Facebook). Their seed round investors include Plug and Play Ventures, Indicator Capital, SV Frontiers and Luc Vincent, the VP of self-driving at Lyft.
