Destro, a startup coming out of stealth, has raised $8 million in seed funding. Unlike many robotics firms, it does not build hardware. Instead, it provides an AI layer that tells robots and human workers what to do in logistics settings. Founder Manthan Pawar says the company's edge comes from understanding customer problems deeply, not from designing robots.
The company's first deployment is with Yusen Logistics, where it runs a pilot using three cart-moving robots from Miva Robotics. The system relies on open-weight vision-language-action models to direct unloading and sorting tasks. After a successful initial pilot, Destro is expanding to 26 robots at one site and 17 at another in Southern California.
Yusen's automation director said Destro adapted its tool for cross-docking, a need that other robot startups did not address. The system replaces paper-based processes with systematic direction, making operations less labor-intensive. Destro plans to replicate this workflow across thousands of warehouses that handle similar work.
Potential limits remain in tasks that require dexterity, where generic robot bodies and open-source models are not yet capable. Pawar is confident the software layer will capture value, though the article notes uncertainty about whether foundation-model developers might eventually compete. For now, Yusen sees no clear use case for humanoid robots, though it is evaluating a pilot with a wheeled humanoid.