Mind Robotics just raised $500 million at a $2 billion valuation to tell the robotics industry it has been thinking about this all wrong.
The company, spun out of electric vehicle maker Rivian in November 2025, is not building humanoids. It is building robots around the physics of the task, not the geometry of the human body — and it has recruited two of venture capital's most aggressive checks to back the argument. Accel and Andreessen Horowitz co-led the Series A, expected to close later in March 2026, one of the largest such rounds in robotics history. Mind previously raised a $115 million seed led by Eclipse Capital, bringing total fundraising to $615 million since founding. TechCrunch Reuters
"Doing cartwheels does not create value in manufacturing," RJ Scaringe, Mind Robotics' founder and Rivian's CEO, told The Wall Street Journal. TechCrunch
That is a direct shot at the humanoid wave — and the timing is not accidental. Figure AI's fleet of Figure 02 humanoid robots was retired from BMW's Spartanburg, South Carolina plant in November 2025 after an 11-month deployment. The robots had contributed to 30,000 vehicles at the plant, according to Figure, but hardware failures in the forearm and cabling made sustained 10-hour shifts untenable. Humanoids Daily Interesting Engineering
Scaringe's argument is not philosophical. It is architectural. "The work happens with the hands," he told TechCrunch at SXSW. "Everything else, from a robotic system point of view, is to get the hands to the right place." The assumption that human biomechanics represent the optimal factory form factor "misses the fact that we didn't evolve in a plant." TechCrunch Advanced robotics, he told Robotics and Automation News, "is going to be critical for global competitiveness, as well as addressing the substantial industrial labour shortages that exist today." Robotics and Automation News
The structural gap Mind is trying to fill is not theoretical. Existing industrial robots excel at repeatable, dimensionally-stable tasks in controlled environments. But a large share of factory value-add work — wiring harnesses that never present themselves the same way twice, materials that flex, parts that shift — requires physical reasoning: the ability to adapt to variability, apply dexterity, and make real-time decisions in dynamic settings. Classical robotics, running on fixed motion plans that repeat consecutively, cannot close that gap. "Wiring harnesses do not present themselves the same way twice; materials flex; parts shift," as a16z put it in its investment thesis. a16z TechCrunch
Mind's answer is full-stack: a foundation model — Scaringe calls it "a brain for how to operate inside of a plant" — paired with purpose-built hardware and the deployment infrastructure to run it. Eletric-Vehicles.com By controlling both the robot and the industrial processes it inhabits, Mind can optimize tasks for robotic efficiency rather than forcing a machine to move exactly like a human. Humanoids Daily The company was originally called Project Synapse — "the name links to synapse and the brain," Scaringe said. He nearly called it Synapse before settling on Mind Robotics. TechCrunch
The founding team comes from Physical Intelligence, Waymo, Zoox, and Google — deep learning and mechanical control, not humanoid bipedalism. The board includes Scaringe, Jiten Behl of Eclipse Capital, Accel partner Sameer Gandhi (who joined in this round), and a Rivian board representative. TechCrunch Reuters
Scaringe's manufacturing credibility is the core bet. a16z General Partner Sarah Wang described him as "one of the very few founders who have built and scaled a vertically integrated hardware company" — someone who built Rivian's full stack from vehicle architecture through electronics, battery systems, embedded software, manufacturing processes, and supply chains. a16z The implied contrast, which Scaringe has made explicit: he is not going to bet Rivian's manufacturing future on robotics companies that have never industrialized a product, lack data flywheels for training models, or lack supply chains. TechCrunch
Rivian's role as partner and major shareholder gives Mind something most robotics startups spend years trying to earn: a live training environment. Rivian is providing data for model training and a deployment floor to iterate in — a captured distribution flywheel that does not require convincing third-party manufacturers to open their factory floors. Reuters Scaringe kicked off the Mind project roughly two years ago, when planning for the Rivian R2 — which costs roughly half what the flagship R1 costs to build — made it clear that building four or five new plants over the next decade required automation no existing vendor could provide. TechCrunch
The timing of the BMW Figure 02 retirement makes Mind's anti-hype pitch look prescient by contrast. But it also raises the hardest question in the story: Mind has told The Wall Street Journal it will have a large number of robots deployed by the end of 2026. TechCrunch The company was founded in November 2025. The Figure 02 fleet took 11 months to fail. Scaringe's deployment timeline is either a genuine inflection point or the most expensive press release in the industrial robotics space, and the data does not exist yet to know which.
Industry experts and investors alike have cautioned that commercializing advanced robotics remains a long road. Training the foundation models that power these systems demands vast amounts of real-world data, and the physical world is unforgiving in ways code is not. "Hardware does not forgive hallucinations," as a16z put it. "Micron tolerances do not care about confidence intervals." Reuters a16z
The market opportunity is real. The industrial robotics market stood at $30.71 billion in 2026 and is projected to reach $93.31 billion by 2035 at 13.21% CAGR, according to Precedence Research — and more than half of production costs across that market still come from labor, making the automation case structurally compelling. TAM Radar Accel Mind is not the only bet in this cycle. Apptronik closed a $935 million Series A in February 2026 at a valuation exceeding $5 billion. TAM Radar Both companies are chasing the same structural opportunity: the moment when AI capability finally meets the physical world's tolerance for error. The difference is architectural philosophy — and the next 18 months of deployment data will answer whether Scaringe's bet against the humanoid form factor was the right one, or whether the market simply rewards whoever ships first.
Scaringe is explicit about what he is not building: "We are not building single-task machines." Mind Robotics The question is whether a full-stack platform built around task optimization, rather than human mimicry, can generalize across the chaos of a real factory floor — and whether Rivian's captured data flywheel is wide enough to get there before the next Figure 02 story breaks.
