For those who haven’t heard, Rivian founder and CEO RJ Scaringe has one other firm — his third by our rely. And this time it’s centered on robotics.
The serial entrepreneur is greatest recognized for Rivian, the buzzy EV maker that was thrust into the highlight in 2018 after debuting an electrical truck with its now-famous gear tunnel.
The publicly traded firm is now just a few months away from delivering its most essential EV but, a mid-sized SUV referred to as the R2. This EV is cheaper to construct than its flagship R1 counterpart (Scaringe tells me “roughly half” the price), and but, in some ways is extra technologically succesful. It’s going to even be cheaper for consumers; the primary model of the R2 will begin at $57,990, 20% cheaper than its base R1T pickup.
After which there’s Additionally, the micromobility startup that started as a skunkworks program inside Rivian and spun out final 12 months backed by VC cash. Additionally, which Rivian maintains a minority stake in, launched a pedal-assist modular electrical bike and cargo quad automobile final October. Scaringe is deeply concerned with that startup as effectively.
Now, Scaringe has robotics on his thoughts. And he aptly named his new startup Thoughts Robotics, which I realized was not its authentic identify. (Extra on that under.)
Thoughts Robotics was not born inside the partitions of Rivian, though the automaker could sometime be a buyer. Thoughts Robotics is a personal firm based by Scaringe that just lately raised a $500 million Collection A spherical co-led by enterprise corporations Accel and Andreessen Horowitz. The corporate, which has raised $615 million since its founding in November 2025, is now valued at about $2 billion.
I sat down with Scaringe on the sidelines of South by Southwest, of which his firm is the first sponsor, to speak about every part he’s engaged on. The Q&A under, which has been edited for brevity and readability, focuses on our dialogue about robotics and the way his new startup is taking a brand new method.
Techcrunch occasion
San Francisco, CA
|
October 13-15, 2026
Are you completed spinning off firms?
Three firms? (pause) In all probability.
Are you positive about that?
Yeah. I imply, yeah. (nods in settlement) Additionally [the ebike company] is right here, in order that’s cool. It’s good to see. It’s cool to see the manufacturers co-existing, which makes me glad.
Thoughts [Robotics] is one thing I’ve been engaged on for some time. Additionally was distinctive as a result of it was constructed inside Rivian, and that we spun it out. Thoughts … I began a complete new firm, Rivian’s like a accomplice in it. I’m wildly bullish on it. It actually advantages Riven, however I feel it has the potential to be a really giant enterprise, simply given the dimensions of what’s occurring.
When did you begin engaged on Thoughts?
So, Thoughts is an attention-grabbing story and it really comes again to R2.
About two years in the past, we had been constructing lots of confidence in R2. So if we’re assured in R2 which means we expect there’s gonna be lots of quantity. If we expect there’s gonna be lots of quantity, it means we expect we’re going to wish lots of crops [factories].
And so successfully, I stated, ‘Boy, if we’re gonna must construct 4 or 5 crops over the following decade, which means we’re going to spend many, many billion {dollars} in capex. What are these crops going to seem like? And we don’t need to construct a bunch of crops after which have them instantly be outdated.’
I kicked off an effort to review the way forward for manufacturing, and basically got here to the view that basic industrial robotics goes to live on like what you see in our plant immediately, or in a Tesla plant or a Ford plant. However the thought of robotics with human-like expertise goes to be actually essential.
We already had a powerful thesis round AI within the bodily world. I met with basically each firm that’s engaged on some type of robotics that may do human-like expertise.
I broadly characterize it as there’s firms that match into the class of present industrial robotics firms. And I got here to the view that they’re going to live on, however not be the varieties of firms that construct robots that may do human-like duties.
After which there’s a brand new batch of firms which can be all like startups, successfully. Most of that are centered on the house, surprisingly — which we may spend hours on. A really giant variety of these firms are centered on folding towels and doing dishes, a subset of these that are industrial.
We simply got here to the view that the substances to achieve success, none of them had all of these substances. And I felt I may construct a greater firm.
I’m not going to construct Rivian’s future manufacturing dependency on firms which have by no means industrialized a product or understanding or have actually frolicked understanding industrial operations, or that don’t have a knowledge flywheel for coaching fashions, or that don’t have provide chains as a result of all these firms, we would want to go assist them discover ways to be an actual firm.
I made the choice to begin a brand new firm, after which had to determine how the governing construction goes to be like.
And on the day I launched the corporate, I raised a seed spherical, after which we simply did this Collection A financing.
While you initially based it or registered, was it referred to as one thing else?
What was the mission identify? I’ll inform you, no person’s requested me that. (pauses) So it was “Venture Synapse.”
Why?
Loads of causes. It is a nice query, nobody has requested this. So the identify of the corporate Thoughts, it type of hyperlinks to synapse and hyperlinks to the mind.
Like firing and misfiring.
Ha, yeah precisely. And it was each due to the mind hyperlink. After which on the time, my youngsters had been in a faculty referred to as Synapse, and I used to be impressed by my youngsters and excited about the enterprise, so I linked them. I really virtually referred to as the corporate Synapse, but it surely’s simply too humorous of a phrase; I needed one thing less complicated, however yeah, it was virtually synapse.
It’s humorous, certainly one of my board members simply wrote me a word and nonetheless referred to it as Venture Synapse.
Who’s in your board?
Myself, Jiten [Behl] of Eclipse, Sameer from Accel, and Rivian has a board consultant. It’s a personal firm, so it’s a lot simpler than a public firm.
You talked in regards to the completely different classes, is what you’re planning on doing most carefully associated to what Boston Dynamics is engaged on, which is the Atlas humanoid, however pairing up with an AI Lab [Google DeepMind]?
We’re constructing the fashions as effectively. We’re constructing the fashions, the robotics, and the infrastructure to deploy at scale.
Then on mechatronics and the robots itself, there was a shocking emphasis on mimicking human biomechanics, or in some circumstances, even going additional, making much more advanced mechatronics.
I feel what’s missed in industrial [robotics] and this is without doubt one of the issues we actually see clearly, is the work occurs with the fingers. So, the fingers are very, crucial. The whole lot else, from a robotic system standpoint, is to get the fingers to the fitting place. And so the flexibility for the robots to do actually advanced motions, like, let’s say, like a again flip that’s really simply means the robotic has lots of pointless complexity in it for the overwhelming majority of duties.
And I perceive the aim of displaying that’s to indicate the flexibleness and functionality (of the humanoid robots). However for those who had been to enter a Rivian facility, you will notice only a few people who have the kind of flexibility that may allow them to do a again flip. And so it’s simply when you consider deploying at scale, you need to reduce the complexity, reduce the variety of failure modes, cut back energy consumption.
And so I do assume lots of the robotic programs, significantly humanoid programs, which can be being thought of for manufacturing are method too advanced relative to what they do. They’re going to be human like, they’ll have fingers, there’s a notion mannequin on the high. There’s going to be the flexibility to regulate and X, Y and Z; there’s going to be locomotion to permit it to maneuver. However I feel, mimicking human biomechanics in a producing setting misses among the basic factors of producing, which is, it’s all hand primarily based.
I’m making an attempt to visualise what it will finally seem like or what the completely different variants will probably be.
So, the fingers are probably the most sophisticated a part of robots, although.
(Rapidly leaping in) However it’s additionally the place you focus your vitality, your {dollars}, the spending, every part ought to go into the fingers.
And I feel the opposite factor to note, for those who’re constructing a enterprise that’s going to service lots of completely different industrial sectors for manufacturing, there’s not one set of fingers that’s going to be good.
If you consider the problem of fingers is torque amplification. For those who’re choosing up a big four-inch diameter metal pipe, it’s a unique gripping resolution than threading an M4 fastener into an aluminum casting. It’s simply very completely different expertise and really completely different ranges of dexterity.
One of many different issues that’s occurring is in robotics, which may be very shocking, is within the organic world we so clearly acknowledge that evolution has allowed us to create, like very completely different optimum units of biomechanics for various issues. The perfect factor for swimming appears to be like very completely different than one of the best factor for operating and one of the best factor for climbing appears to be like completely different than one of the best factor for lifting. And due to that, people are usually not significantly good at swimming. Like, for those who had been to place me up in opposition to a dolphin, I might lose. For those who put me up in opposition to a cheetah, I might lose.
We’re extra highly effective, in fact, in our mind. And so in manufacturing, to imagine that the form and type issue of a human is the optimum form and type issue for every type of labor, I feel, simply misses the truth that we didn’t evolve in a plant. Now, crops have advanced round us, which is a crucial level. And there’s an enormous embedded brown subject infrastructure that robots must plug into. However the half that they must plug into is that they must roughly match within the area that people permit for, they usually must have fingers. I name that every one out as a result of we’ve a unique perspective then I feel lots of the robots that always began from saying we’re going to work within the residence, which is a really completely different ODD [operational design domain].
However like all these use circumstances that say we’re gonna do every part a human does. it’s a unique ODD and subsequently it’s gonna result in a unique type issue than for those who stated, I’m simply gonna be centered on being a producer.
And truly [in an industrial factory] you don’t must cope with stairs, you don’t have carpet to tile transitions, you don’t have to fret about tripping on a cat. You don’t have to fret about stepping on a child. So that you simply have far more pleasant constraints for deployment at scale. And you may map the setting, and the setting doesn’t change, proper?
The one factor that may change the dynamic could be any interplay with people [in a factory].
Which is a crucial level.
That’s why the UI [user interface] is essential. We haven’t proven what it’s gonna seem like, however I’ll provide the transient [sic]: It must really feel pleasant, as a result of it’s gonna work a very long time with people. However it doesn’t need to really feel dopey and find yourself with a kick me sticker on its again, proper?
Loads of robotics has additionally gone in direction of these very lean, muscular types that look very Terminator like. I don’t know why. I suppose science fiction films or one thing has brought about the economic design world to try this. There’s a model of robots they usually all type of mix collectively. They’re these very athletic sprinter physique types. I feel there’s a unique type issue that may really feel far more approachable, however not dopey.
I’m beginning to see a thread with what you probably did with Additionally, and your considering, when it comes to modularity with Thoughts.
A few of that, yeah.
The opposite factor is Rivian. For those who had been to ask any person in 2018 what a 1,000 horsepower truck appears to be like like, and also you squinted, you wouldn’t consider an R1 [truck]. And so, I feel the thought is you’ll be able to have all this functionality, but it surely doesn’t must look intimidating or scary. It may be pleasant — it’s a design perception and aesthetic that can carry into robotics as effectively.
