Startup Destro AI has emerged from stealth with an $8 million seed round led by Base10 Partners and Bonfire Ventures, with participation from CoFound Partners. The company is betting that improving robot operations inside warehouses may be more valuable than manufacturing a new robot, as it builds a software layer that coordinates robots, workers, and vehicles within a single workflow.
The company is led by Manthan Pawar, who holds a master’s degree in robotics from NYU Tandon and has around eight years of experience in supply chains and robotics in the United States. According to comments he made to TechCrunch, Destro is targeting specific operational problems in warehouses and says it is on track to achieve positive cash flow by the end of the year.
A Platform That Coordinates the Entire Operation
Destro is currently focusing on cross-docking operations, which involve unloading goods from a truck and sorting them into mixed loads that are later moved onto other trucks for final delivery. The company began a pilot at a Yusen Logistics facility in the Pacific Northwest, using three robots manufactured by Miva Robotics to move carts.
The robots operate through the Vision system, an operating system based on open-weight vision-language-action models capable of interpreting camera images and instructions and controlling the robot’s movement. At the same time, the Mothership system manages the broader operation; when workers unload goods into different carts, the robots identify completed carts and move them to their destinations.
Yusen says this approach made the process more systematic and less dependent on manual labor and paper documents. The key distinction is that Destro does not merely move carts from one point to another; it directs the worker, cart, and trucks within a shared workflow.
What Changed in Practice?
According to Richard Brunelle, automation director for Yusen’s U.S. logistics services division, solutions from two large robotics companies were tested before Destro was selected. One solution was able to move carts but did not manage loading and unloading operations, while the other provided fleet-management tools but left overall coordination to a person. Destro was selected because it adapted to the cross-dock environment and provided autonomous direction for the entire operation.
Yusen is now expanding the initial pilot into a full deployment involving 26 robots, while also beginning a second pilot using 17 robots at a facility in Southern California. Pawar intends to replicate the same workflow in thousands of warehouses that carry out similar operations.
Limitations and Open Questions
The success of the current model does not mean that the platform is suitable for every warehouse task. Work requiring manual dexterity and precise object manipulation remains a challenge for general-purpose robots and open models. It is also unclear whether companies developing foundation models, highly dexterous robotic hands, or general-purpose humanoid robots will make their tools available to software-layer companies or compete with them directly.
Destro’s experience reveals a practical trend in logistics automation: the value does not necessarily come from the robot’s form, but from integrating the robot into a complete operational process. However, the repeatability of this model will depend on the similarity of workflows across warehouses and on the systems’ ability to handle tasks that go beyond transport and routing to precise grasping and manipulation.