Last week, US regulators banned a wide variety of Chinese-made robots from being imported to the United States because of concerns about national security. Saurabh Chandra was ready to seize the moment. His company, Ati Robotics, assembles its robots in India, and Chandra believes they use fewer Chinese parts than almost any other robotics company in the US.
Investment in robotics startups hit record highs this year, both in the number of deals and dollars invested, according to PitchBook, which tracks venture capital flows. But many of those companies primarily develop software, and they largely still rely on Chinese suppliers for their hardware.
The Federal Communications Commission’s decision to ban new models of human-like robots, known as humanoids, and other advanced devices from China could hamper the growth of many smaller robotics firms in the US. But some businesses could benefit, including Ati.
The startup was founded in 2017 to develop motors for self-driving cars. Years later, it zeroed in on making its own robots, including tuggers and pallet movers that cart tons of materials throughout big facilities. Ati developed its own hardware, believing that it would lead to better functionality, a bet that Chandra says many of his advisers cautioned against but one that wound up naturally limiting reliance on China.
It has begun to pay off. Ati has several hundred robots currently in operation at warehouses and factories and more than 50 customers in total, Chandra says. Ati’s first humanoid, which moves heavy bins around, will go into service later this year.
Chandra spoke with WIRED last week about how Ati Robotics steered clear of China, while building robots he argues can still compete with the industry’s best on price and performance.
This interview has been edited for clarity and length.
WIRED: Developing hardware is expensive. How were you able to financially sustain doing that?
Saurabh Chandra: R&D is in Bangalore, which helps with our labor costs. Bangalore was undergoing this electric vehicle revolution in the two-wheeler, three-wheeler category in India. And it turns out that the kind of vehicles we were making for the robotics industry happened to have a similar power profile and characteristic as the two-wheeler and three-wheelers being made in the EV industry.
We tried to leverage the components and supply chain being developed for those vehicles. Typically, automotive parts are very reliable and are produced in large volumes, hence they are very good and cost-efficient. Though we had to do engineering with some of those vendors to make the parts work in a robot. But once we did, we had a great win.
One of your robot models is China-free, right, but not the others?
They have different levels of what I might call “supply chain resiliency.” Two of our best-selling models—our 10,000-pound tugger and the pallet mover—use very few parts from China. They’re tugging trolleys—you may be taking raw material from stores to the factory line or from one process point to another.
But even with the humanoid that we have, the full arm is done here in Bangalore. The actuator is also done here, but we do get the harmonic drive from China, for example, and the frameless motor, at least for now.
The battery cells come from China. But we could, for example, get the cells from South Korea or Japan tomorrow. So those are passive components. There are some components that do come from China when we go deeper. We use induction motors, so even the magnet in the motor is not coming from China. That’s a very different choice from many of the people who are doing robots.
Which of your robots uses the fewest parts from China?
The 10,000-pound tugger. It’s really China-free actually. And even the pallet mover has maybe less than 5 percent Chinese parts.
Which part of the tugger do you still need from China?
Just the cells in the battery. In the pallet mover, we use one very good solid-state lidar sensor, which we don’t have an option from outside China. But in the tugger, we use Ouster, which is a US-based lidar sensor.
Did you have to make some trade-offs?
Yeah, we used induction motors, which are slightly less efficient than, say, Chinese-made brushless motors, which have magnets in them. But you can trade off a bit of efficiency for a bit more battery.
How do you market your robots? Is your supply chain part of the appeal?
That’s not been a big part of our marketing, to be honest, but we do let customers know. We have not seen this being a very explicit part of their buying criteria. But many customers have definitely told us that it influences their decisions.
How do you plan to make your new humanoid robot less reliant on parts from China?
It’s very early. We don’t try to optimize the supply chain from the get-go. The gears come from China right now, but those gears are also manufactured in Germany, Taiwan, and Japan. So they’ll be a little more expensive than the ones we get from China, but alternatives are available.
With this new ban in place, do you think it’s going to be easy for US companies to find alternative suppliers?
The regulation is not saying that you cannot use any part from China. But quite a few people will be impacted in my opinion. Absolutely. Most of these robots have been invented in the US. It’s a bit of a tragedy that we are not making them in the US. Maybe more people will be encouraged to make them locally, which is not a bad thing.
Where is your humanoid ultimately assembled?
Currently, we manufacture all of these robots in Bangalore, but we are setting up a facility in Madison Heights near Detroit, Michigan, where we will eventually assemble them in the future.
How are people in the robotics industry reacting to the new ban?
Everyone’s in denial. Absolutely. It’ll get real. A couple of months ago, I had met people at a conference, and they were actively trying to lobby in Washington, DC, to prevent something like this from happening, but I don’t think they succeeded.
Who do you think will ultimately benefit from the new ban?
Everybody who is trying to make robots outside of China, including robotics companies in the US and Europe. We are an accidental beneficiary. Tesla’s Optimus is obviously a clear beneficiary. Figure (a robotics startup) manufactures in the US. I hope more people get inspired. It’s not like what we are doing is magic.
