Destro, whose startup emerged from stealth with an $8 million seed round on Tuesday, says the best path to successful robot deployment is to not build robots at all.
Most emerging robotics companies are focused on finding the right form factor or developing AI models that make metal hands work. Instead, Destro built an AI intelligence layer to help robots work efficiently on the logistics floor by telling human workers what to do.
“A lot of robotics companies started with robotics engineers asking, ‘What great things can we do?'” industry veteran Manthan Pawar, founder of Destro, told TechCrunch. “One of the biggest reasons we can beat robot companies is because we are not a robot company.”
“Our thinking is that we know our customers’ problems very well and are actually on track to be cash flow positive by the end of this year,” explains Pawar, who earned a master’s degree in robotics from NYU Tandon College and spent about eight years in the U.S. supply chain and robotics industry.
Richard Brunell, director of automation for South Korean shipping giant Yusen Logistics’ U.S. logistics group, helped steer the startup in the right direction. Brunel is responsible for automating about 30 facilities in the United States. So Yusen uses fixed automation such as conveyors and sorters, but it also continues to pilot new technologies such as trailer unloading robots and autonomous floor scrubbers.
Finding more ways to leverage technology like this is a priority for Brunel. Yusen’s British subsidiary has built a fully autonomous facility, and similar plans are currently underway in the United States. When Brunel met Pawar, Destro was focused on another challenge in the field: picking and packing, the process of sorting different products into a single package.
Brunel realized that his main challenge was “fundamentally the same problem.” His work included cross-docking. This meant unloading goods from one truck, sorting them into mixed loads on other trucks, and finally delivering them to the customer.
“I asked if Destro was willing to adapt their solution for that cross-dock environment, because I don’t think anyone was doing that,” he told TechCrunch. “They walked away, thought things through, and came back excited. They agreed that no one else was working on that area, and that they could change the tools to support it. They’ve been going all out ever since.”
Destro began with a pilot program at Yusen’s facilities in the Pacific Northwest using three cart-mobile robots manufactured by Miva Robotics and operated by Destro’s Vision operating system, which is based on Openweight’s visual, verbal, and behavioral model, an AI system that interprets camera images and commands to control the robot’s movements.
As human workers unload goods from one truck to another, a robot finds a full load and transports it to its destination. In Brunel and Pawar’s assessment, this is efficient because the humans, carts, and trucks are all controlled by Destro’s mothership operating system.
“We can reduce the effort of doing that and move away from newspapers,” Brunel said. “Now everything is systematic.”
This is important because Yusen also spoke to two other major robotics startups in this space, but neither could fit into their workflow. One company offered a robot that could move carts point-to-point, but could not manage loading and unloading. The other one came with fleet management tools, but needed someone to orchestrate everything. Destro won because of its ability to apply autonomous direction throughout the process.
Destro has now expanded its first pilot to a full deployment of 26 robots and is launching another pilot with 17 robots at Yusen’s facility in Southern California. Pawar plans to adopt this cross-dock loading workflow and “copy and paste” it to thousands of other warehouses that handle this type of labor. It was this plan that convinced Base10 Partners and Bonfire Ventures to lead Destro’s seed round with additional investment from CoFound Partners.
Destro can face challenges with workflows that require more dexterity and manipulation. Common robot bodies and open-source models have not yet proven capable of these tasks, and it is unclear whether companies building basic models, highly dexterous hands, or general-purpose humanoids will make those tools available to companies like Destro, rather than trying to capture that business themselves.
Pawar remains confident that he has found a place in the stack where value can be created and that his company can leverage the research funding poured into new AI models and robotic components.
“The robot is a platform. Every layer model is a platform. But we build a harness around it,” Pawar told TechCrunch. “That harness is so complex and value-added that these workflows would be completely impossible without it, right?”
For now, competition from future robot manufacturers is probably not an issue. Brunel said he has not yet found a suitable use case for a bipedal humanoid robot, although he is evaluating potential pilots for wheeled humanoids.
“At the end of the day, we’re not putting automation into buildings just because the technology is interesting,” Brunel said. “We need to solve real operational problems. If we can make operations more efficient, consistent, and flexible, our customers will ultimately benefit.”
Destro will be making the case for robot companies that aren’t robots at TechCrunch Disrupt in just a few weeks. Join us in October to see Destro first-hand along with the rest of Battlefield’s competitors.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
