LAS VEGAS — At Manifest 2025 this week, several exhibitors are focusing on methods for collecting, analyzing, and using supply chain data. Gather AI yesterday announced that it will augment its DJI drones with new U.S.-made Starling 2 Logis drones powered by the ModalAI VOXL 2 autopilot for gathering inventory data in customer warehouses.
“Our hardware-agnostic AI and computer vision stack empowers logistics and supply chain customers with unparalleled flexibility and access to cutting-edge solutions,” stated Sankalp Arora, Ph.D., co-founder and CEO of Gather AI.
”Our inherent ability to upgrade to better hardware has enabled us to launch a more capable and efficient inventory intelligence device, which is completely made in the USA,” he added. “You will see Gather AI’s AI stack support more forms of hardware as we pursue pervasive digitization of warehouse workflows.”
Smarter software and a U.S.-produced drone platform can address user concerns about national security, supply chain bottlenecks, and labor shortages, noted Arora.
Starling 2 Logis to increase inventory accuracy
The new drone inventory system is twice as fast as previous systems, and it can improve accuracy with 3D data, Arora told Automated Warehouse. This enables warehouse associates to focus on exceptions, he said.
“There’s plenty of robotics hardware for moving goods in the warehouse, and digitalization might help with the unsolved problems of picking,” said Arora. “But operators can get an immediate ROI [return on investment] from inventory visibility.”
Gather AI listed the following benefits of using the Starling 2 Logis drones:
- They can scan 25 times faster than manual methods, it said.
- The drones increase the inventory-scanning speed from an average 900 pallets per hour up to 1,500 pallets per hour
- They can increase inventory accuracy to 99.9%, the company claimed.
- They can also fly in narrower aisles. DJI drones can fly in very narrow aisles (VNAs, typically as narrow as 5 ft., 4 in. or 1.6 m wide), but Starling 2 Logis drones can fly in aisles as narrow as 4 ft., 5 in. (1.3 m) wide.
- The drones have new dynamic obstacle-avoidance capabilities.
- The Starling 2 Logis offers richer inferred case counting and location occupancy estimation capabilities, thanks to image and depth data access, especially in double deep racking contexts.
Gather AI interprets ‘beyond the barcode’
Three Carnegie Mellon University alumni founded Gather AI in 2017. The Pittsburgh-based company said its computer vision technology enables drones to fly autonomously without any GPS, Wi-Fi, or infrastructure changes.
Gather AI said its machine learning algorithm analyzes inventory pictures and can read and interpret data “beyond the barcode,” including lot codes, text, expiration dates, case counts, and occupancy information. Facility operators can compare their real-time physical inventory with data from the warehouse management system (WMS) to make changes as needed to maximize accuracy.
“Key enablers include cameras getting ubiquitous and hardware sufficient for digitizing environments,” Arora told Automated Warehouse. “A good analogy would be that you can’t have Uber without smartphones. You can’t have AI without these cameras and IoT [the Internet of Things].”
Users of Gather AI’s systems include third-party logistics providers (3PL), retail distributors, manufacturers, and food and beverage companies. With its software, the new drones will enable warehouse operators further improve inventory efficiency, said the company.
“Now, with generative AI and large language models [LLMs], developers can rapidly iterate, and users can run natural-language queries,” explained Arora. “With agentic AI, there’s the promise of getting systems to self-improve and affect the real world.”
Gather AI’s customers include GEODIS, Langham Logistics, NFI Industries, Barrett Distribution, and the Army & Air Force Exchange Service (AAFES). Last month, it offered a partner program to better reach and support clients.
“We worry about being the best, not necessarily the first,” said Arora. “We happened to be among the first to go into freezers, because we’re hardware-agnostic.”
Save now with early bird registration
ModalAI designs VOXL to speed time to market
“ModalAI is proud to partner with Gather AI on the development of the Starling 2 Logis,” said Chad Sweet, co-founder and CEO of ModalAI. “Gather AI’s rapid time to market is another win for ModalAI and our mission to help developers accelerate their path to autonomy with our U.S.-made VOXL 2 autopilot, ready-to-fly development drones, and open software.”
The San Diego, Calif.-based company said its SWAP-optimized Blue UAS Framework VOXL autopilots integrate AI modules for retail, security, government, and home applications. The company said its modules can help industrial users design autonomous aerial and ground-based systems to navigate in GPS-denied environments.
ModalAI asserted that its VOXL product line can help manufacturers and independent builders build smaller, smarter, and safer drones and robots and get them to market quickly and affordably.
When will the convergence of robots and drone hardware with artificial intelligence and data analytics — or physical AI — be as ubiquitous as digital cameras and smartphones? Gather AI is already working with early adopters, replied Arora.
“There’s a stark difference between those early adopters and the long tail of laggards,” he observed. “But the software will help them catch up over the next two to three years.”
