How Mbodi Is Solving Robotics’ Scaling Problem with Xavier Chi

In this talk

1. 2:20

   Making robots easier to deploy

   Mbodi builds AI software that lets people teach industrial and collaborative robots new tasks using natural language. Chi says traditional automation often requires extensive integration, fencing, and custom programming, leaving a robot able to perform one task on one line. He described manual picking work at a pharmaceutical plant as an example of a task that could be automated.

2. 4:39

   Finding customers and partners

   Mbodi’s current use cases include pick-and-place, packing, and kitting, where products or tasks can vary from week to week. Chi said the company works with ABB Robotics after winning its AI Startup Challenge in 2024, and also pursues direct sales. He pointed to labor shortages and one warehouse customer that overhires by 20% because workers may not show up.

3. 11:04

   Why robotics software is hard to scale

   Chi described robotics sales as involving software companies, robot makers, system integrators, and end customers. Integrators may focus on large warehouse projects where robotics is only a small part, while selling deployments directly can require costly customization for each customer. He sees generative AI as a way for robots to adapt to tasks with less customization.

4. 15:13

   Funding and proving reliability

   Mbodi started with a small pre-seed round, won the ABB challenge, joined Y Combinator’s Spring 2025 batch, and raised a seed round afterward. Chi said founders need to show that a customer problem is real, their solution works, and the team can scale it. For reliability, Mbodi runs systems for extended periods and reported 99.6% success over an eight-hour test.

5. 17:44

   Building through real deployments

   Chi favors deploying practical robot solutions over waiting for a single model that can perform every task. He said robots must improve across speed, cost, and flexibility, with progress in both software and hardware. Mbodi’s near-term focus is to deploy in warehouses and factories, use production data to improve its system, and hire robotics and AI engineers.