Rivian‘s founder and CEO RJ Scaringe has shared new details about the company’s AI-focused spin-off Mind Robotics.
The Irvine-based company announced late last year the creation of its second standalone firm — after the micromobility spin-off ALSO. — with an initial $115 million in external seed capital.
Last week, Scaringe revealed that Mind Robotics had secured an additional $500 million in funding bringing the startup’s valuation to $2 billion.
In an interview with Inc. published on Tuesday, Scaringe explained the company’s purpose is “to think about robotics through the lens of creating human-like capabilities for industrial applications.”
The founder of the company said he is “incredibly excited and bullish on this,” adding that the first Mind Robotics robots are already performing human-like tasks at Rivian’s facilities.
Rivian provides the company with a large data flywheel for training its models, as well as an at-scale launch environment, as previously stated by the robotics firm.
Upon Mind Robotics‘ launch, Rivian said that its spin-off company would focus on “the advancement of industrial AI to reshape how physical world businesses operate,” highlighting the “synergies shared between the development of autonomous driving and physical AI.”
Approach to Robotics
Mind Robotics focuses on improving manufacturing efficiency, Scaringe said on Tuesday.
“We’re building both a foundation model, so a brain for how to operate inside of a plant, but also the mechatronic embodiments,” he said, “so the robotics to then actuate what the brain is capable of understanding.”
Scaringe emphasized that the company aims to address a shortage of manufacturing labor in the United States.
“I think this is really how we both lower the cost of goods by making manufacturing a lot cheaper,” he said, adding that it will “also allow manufacturing to really become a strength in the Western world, where we don’t have labor availability and cost of labor is also very hard to compete globally here.”
Earlier this week, discussing the same topic with Fast Company, the Rivian CEO also acknowledged this extreme shortage of industrial labor in the Western world, stating that it poses a challenge to scale plans.
“We often talk about bringing manufacturing back to the United States. One of the challenges we have, we don’t have people to work in manufacturing in the United States,” he said then. “And if we plan to scale, which of course we do, we’ll need to solve this.”
Rivian‘s Chief has repeatedly pointed out that Chinese competitors benefit from lower labor costs (alongside national subsidies), while also acknowledging tech advancements from companies such as BYD and Xiaomi.
In the West, however, “there’s not a workforce who wants to do that. The cost structure can’t support being globally competitive.”
Therefore, to Scaringe, the way for the United States to compete is through technology.
“I see a real opportunity to build one of the strongest and most capable industrial robotics companies in the world,” he added.
Manufacturing
According to RJ Scaringe, the current robots used in industrial applications are “very classical,” running on “fixed motion plans” that repeat consecutively.
In his view, these automated processes are “not capable of human-like skill in terms of dealing with variability, having any level of reasoning, or having any level of adaptation based upon circumstances.”
The Rivian CEO distinguished two types of robotics companies.
The first category includes the ones producing traditional six-axis robots, which he thinks “will continue to make those robots” and “be successful for classic industrial applications or industrial use cases.”
However, he doesn’t see those firms playing a role in the next wave of automation.
The second category, he explained, consists of the “new companies that have emerged recently that are building robots that we have more human-like skill.”
After meeting with several of these companies, Scaringe decided to start his own robotics enterprise, citing what he described as “a unique set of ingredients between a huge data flywheel with Rivian, a massive deployment environment with Rivian as our first customer, a need within Rivian.”
Other Form Factors
Late last week, speaking with Matt Berman, Scaringe also touched on the different form factors for robots filling human-like tasks.
He acknowledged that one option is “something that mimics the biomechanics of a human directly, so pure humanoid,” but expressed confidence that “with more imagination, we’re going to see a lot of different robotic form factors.”
When the host questioned him on the current development of humanoid robots — by companies such as Tesla and Chinese competitors — Scaringe recognized that “the investment is right,” but pushed back on the idea that humanoids are the only path forward.
“I’m just calling out that I think there might be more variety. There’s going to be a lot more variety,” he stated.
Tesla is developing the Optimus humanoid robot, for which Musk confirmed earlier this year that production is scheduled to begin this summer.
High-volume production is targeted for 2027, with an Optimus 4 set to follow after that.
In the company’s latest earnings call, Musk said Optimus units are currently working in Tesla factories, but primarily for learning and data collection, rather than performing “useful work.”
The CEO said the robots are “still very much in the R&D phase.”
At Tesla‘s Annual Shareholder Meeting last November, he hinted at use cases like having Optimus serve as a deterrent for crime by monitoring convicted criminals, and even using (his company) Neuralink to upload human consciousness onto the robot.
‘Human-like’ x Humanoid
While acknowledging that there will be use cases for humanoid robots, Scaringe argued that “in manufacturing, if you need arms with hands to do things that humans do with a lot of dexterity, it can look human-like, but it doesn’t need to be exactly like a human.”
To illustrate the point, he turned to self-driving.
One approach, he said, would be to “make a robot that looks exactly like a human and sits in the driver’s seat and actuates the pedal, turns the steering wheel, shifts the car into gear.”
But there is “another way — which I think is a much better way — is to mount perception around the periphery of the vehicle on the outside of the vehicle and have the actuation of, let’s say, the brakes and the controls not have to happen through something that has a lot of lag, which would be your arms and your legs, but happen electronically.”
“I think we can easily see an autonomous vehicle as a robot, but it’s not like we started that by saying, let’s make a human form factor that drives our cars,” he noted.
