Construction is a massive portion of global GDP at 13%, but since 2000, its labor productivity rate has fallen far behind overall economic growth and productivity in sectors like manufacturing. The problem is significant in the U.S., where construction contractors are facing stagnant productivity coupled with a historic labor shortage that began during the Covid pandemic.
Part of the reason is that construction has seen little of the automation that transformed factories and warehouses through robotics and automation. Those are controlled environments where programmed robots can repeat the same task. Construction sites are not. Every site is different and complex. In recent years, sweeping advances seen in physical AI have shown the potential of mass-scale construction robotics.
Origin builds integrated software and hardware that mirrors successful agents in software. The base model is trained on synthetic and real-world construction site data to handle perception and planning.The robot’s arms use a range of power tools, and each tool is trained separately on different tasks at live job sites, the same way AI agents in software learn to use separate tools.
For contractors, Origin’s general-purpose robot is a smart, well-trained worker that only needs high-level instructions like “sand and paint this room.” For workers, it means the robot takes on the most physically demanding and unsafe jobs on site, and does them five times faster.
The Origin founding team is well-equipped to propel the U.S. construction industry forward blending deep business and operations experience with AI and hardware expertise. Yogesh sold his previous company to Amazon and Tushar led core AI teams at Eightfold. Ashish brings a decade of hands-on hardware robotics experience. This blended knowledge is reflected in Origin’s execution pace: curating the largest dataset in the world for indoor construction environments and training manipulation models that outperform every available robotics model by a steep margin. The team moved from concept to live customer deployments in the U.S. all before hitting their Series A.
Yogesh, Tushar, and Ashish bring the business, AI, and hardware experience this problem needs and are ready for the next phase of rapid execution. We look forward to working alongside them on the journey.
