As robotics models improve, one of the biggest opportunities is getting robots deployed and doing economically valuable work. That requires a lot beyond the model’s intelligence itself: hardware that’s easy to install, integration into existing workflows, and support and fallbacks that keeps the robot working when something unexpected happens. That’s the opportunity Ultra is focused on, and why we’re excited to share our investment in the company.
Ultra has built its product around making deployment straightforward. Its Operator robot fits into existing warehouse workstations and can start packing orders within hours of being installed. Customers pay an installation fee and a monthly subscription covering hardware and software support. A warehouse can start with a specific task and one robot and add capacity as it sees an ROI from it.
Starting with e-commerce order packaging gives Ultra a frequent, repetitive job with a clear customer need. It also exposes the robots to variety in the form of different products, flexible packaging, and workflows that change across customers. Ultra’s robots have already packed more than 500,000 orders across US warehouses.
We believe this deployment focus is especially important because of how robots learn. Putting robots to work generates data about the situations they encounter and where they struggle. Ultra works with various model labs such as Physical Intelligence to turn that experience into better model performance. When a robot needs help or fails, a human operator steps in, completing the task. This provides valuable data that models can be post-trained on to do better at these use cases.
Over time, this can become a reinforcing cycle. More deployments means more useful data which results in better models. Over time, better models means less intervention which improves reliability and economics. Building the customer relationships and operating infrastructure to sustain that cycle is a substantial undertaking, and a core part of the opportunity we see in Ultra.
The team brings experience directly relevant to that work. Jon Schwartz, Max Friefeld, and Oliver Ortlieb previously built Layer By Layer and Voodoo Manufacturing together. Chetan Parthiban, Ultra’s fourth co-founder and chief scientist, brings applied machine learning experience from Arena. Their combination of manufacturing, operations, and AI experience is well suited to getting robots working reliably in customer facilities.
We’re proud to partner with Jon, Max, Oliver, Chetan, and the entire Ultra team as they build a best-in-class modern robotics deployment company. If you’re interested in joining them, Ultra is hiring.
