The first robots from Mind Robotics will not try to run an entire Rivian production line. They will start by picking parts out of bins and placing them in fixtures, the kind of repetitive handling that eats up to half the labor in some sections of an auto plant. Rivian CEO RJ Scaringe said in a conversation published by venture firm Kleiner Perkins on November 9 that the machines will handle about 20% of tasks on the factory floor when they arrive at Rivian's Normal, Illinois plant in mid-2027, with a longer-term goal of 80%.
Mind Robotics is the industrial robotics spinoff Scaringe founded last year, built around the idea that conventional factory automation has never been good at fiddly, judgment-based work. Rivian is both a shareholder and an operating partner, giving the startup access to a live, high-volume factory floor to train and test its models. The company has raised more than $1 billion since its founding, including a $400 million round reported in May, and recently hired former Tesla and Ford EV engineering chief Doug Field as president.
Why the 20-to-80 split matters
Scaringe's numbers are a deliberate signal to an industry that has struggled to translate humanoid and industrial robot demos into sustained production work. Pick-and-place tasks are the easiest to automate because they are well-defined and repeatable; the harder 60 percentage points involve recognizing part variation, adjusting grip in real time, and handling exceptions that a fixed program can't anticipate.
The approach Mind Robotics is using leans on video. Engineers film a broad spectrum of plant tasks, feed the footage into software that rates each task for the dexterity and reasoning it requires, and then judge whether the company's launch robot can attempt it. Scaringe said the company has recruited staff from the consumer electronics industry specifically for expertise in tactile sensing, a bet that better touch feedback — not just vision — is what closes the gap between 20% and 80%.
Where Mind Robotics sits among industrial robot makers
Mind Robotics enters a field already crowded with well-funded competitors chasing the same factory-floor dexterity problem, including Figure, Apptronik, and Path Robotics, which just raised $300 million at a $2.45 billion valuation. What differentiates Mind Robotics is its built-in customer: Rivian's own assembly lines serve as both testing ground and first deployment site, a structure similar to how Amazon has used its warehouses to iterate on robotics internally before any external sale.
Rivian's hardware strategy adds another wrinkle. The company's in-house RAP1 chip, originally designed for vehicle compute, is being adapted for use in Mind Robotics' machines, an attempt to reuse automotive silicon investment rather than buy specialized robotics processors from Nvidia or others.
Scaringe framed the economics around headroom: designing hardware capable of running increasingly capable software, so that the same physical robot can climb from 20% to 80% task coverage as its underlying model improves, without a hardware refresh. Whether that scaling curve holds will become testable once the first units land on Rivian's floor in mid-2027.