Mbodi will show how he can train a bot using AI agents at TechCrunch Disrupt 2025

Robots can be programmed to do a variety of tasks, such as packing boxes and even performing surgeries. But each individual movement or task requires its own training process, making it difficult for robots to adapt to real-world scenarios.

Mbodi wants to make training robots easier and faster with the help of AI agents. The company will showcase this technology as one of the top 20 finalists at TechCrunch Disrupt 2025.

New York-based Mbodi has built Cloud to Edge, a hybrid computing system that uses both cloud and on-premises computing, and is designed to integrate with existing robotic technology stacks. The program relies on a large number of AI agents that communicate with each other to collect the information required to help the robot learn the task faster.

Once deployed, Mbodi will collect data and learn from real-world use cases.

Xavier Chi, co-founder and CEO of Mbodi, told TechCrunch that users prompt the software using natural language, and Mbodi breaks the request down into smaller subtasks. Mbodi’s customer group essentially breaks down the task of gathering the information required to instantly train and defeat the robot.

“The difficult thing about the physical world is that it is infinite possibility,” Chi said. “Every time you can invent something completely new, you don’t have any data, and that’s a problem in the physical world. We always need a system where you can coordinate different models or have someone debug the robot and tell it to do certain things in certain ways.”

Chi said he and co-founder Sebastian Peralta got the idea for the company while working as engineers at Google. While they were not working on robots, they both realized that progress in AI was moving into the physical world and despite the rise in physical AI, there was no great way to train robots quickly.

Several companies, such as Skild AI and FieldAI, are looking to help make robot training faster by building large, global AI models that contain enough real-world data to make it easier to adapt to new environments. Che said that philosophy is out of proportion to how much the world is constantly changing.

Mbodi was launched in 2024 with a focus on picking and packaging. The company won the ABB Robotics AI startup competition last year, earning it a partnership with the Swiss robotics organization that SoftBank acquired for $5.4 billion in October.

The company is now working with a Fortune 100 consumer goods and product company on a proof of concept.

“For consumer packaged goods customers, they have a lot of people, and they are packing different products of their brand into a tray or a shelf, and the problem is that it changes every day,” Qi said. “For this reason, it is impossible to put robots there. As for reprogramming these robots, that is not possible. There are still a lot of humans doing this work.”

Mbodi hopes to begin rolling out her programs further in 2026.

“We want to build something that works and can actually be deployed,” Chi said. “We’re not a research lab; we don’t want to be a research lab in that regard. We want to put something into production that works reliably.”

If you want to hear Mbodi live, see dozens of additional performances, attend valuable workshops, and make connections that drive business results, Head here to learn more about this year’s Disruptheld from October 27 to 29 in San Francisco.