A Quiet Side Project Becomes Robotics' Loudest Bet

RJ Scaringe, the founder and CEO of EV maker Rivian, has spent the better part of the last 18 months quietly building a second company. On April 7, 2026, that company — Mind Robotics — emerged from stealth with a $500 million Series A co-led by Accel and Andreessen Horowitz, valuing the Palo Alto–based startup at $2 billion post-money. The round comes barely 14 months after a $115 million seed led by Eclipse Ventures, and lands Mind Robotics squarely in the center of the "physical AI" gold rush that has minted unicorns like Figure, 1X, Skild AI and Physical Intelligence over the past year.

What sets this deal apart is not the size — mega rounds in robotics have become almost routine — but the founder profile and the wedge. Scaringe is one of the very few American hardware founders who has actually shipped a complex consumer product at meaningful volume. Pairing that manufacturing DNA with a foundation-model approach to industrial manipulation is, according to investors, the closest thing the sector has to a de-risked bet.

Company Background: Born on the Rivian Factory Floor

Mind Robotics was incorporated in late 2024 after roughly two years of internal R&D inside Rivian's Normal, Illinois plant. Scaringe has often spoken publicly about how the lack of flexible, AI-native industrial robots was one of the biggest pain points in scaling R1T and R1S production. Rather than continue building tooling internally, the team spun the effort out as a separate company, with Rivian retaining a minority equity stake and licensing its perception and motion-planning IP to the new entity.

Day-to-day operations are run by CEO Peter Chen, a co-founder of Covariant and a UC Berkeley PhD whose previous company was acquired in part by Amazon in 2024. CTO Anand Gopalan, formerly head of perception for Tesla's Optimus program, joined in early 2025. Scaringe serves as executive chairman while remaining CEO of Rivian. The company now employs roughly 280 people, with hires concentrated from Waymo, Nvidia Isaac, Boston Dynamics and Rivian's own automation team.

Its first product, internally codenamed M1, is a wheeled, dual-arm mobile manipulator running a proprietary vision-language-action foundation model called MindOS. Unlike Figure or 1X, Mind Robotics has explicitly rejected the humanoid form factor, arguing that legs add cost and failure modes without delivering meaningful productivity in structured industrial environments.

Inside the Round

The $500 million Series A is co-led by Accel and a16z, with participation from existing investors Eclipse Ventures, Lux Capital, Khosla Ventures and K5 Global, all of whom exercised pro-rata rights. New investors include Nvidia's NVentures, Sequoia Capital, Thrive Capital, T. Rowe Price, and Sanabil Investments (a subsidiary of Saudi Arabia's PIF). Siemens joined as a strategic investor, taking a small equity stake alongside a commercial agreement to integrate MindOS with Siemens' factory automation stack.

The $2 billion post-money valuation represents roughly a 5x markup from the seed and is, by pre-revenue robotics standards, aggressive. It is, however, dwarfed by the ~$39 billion that Figure reportedly commanded in its 2025 round, suggesting investors are paying a discount in exchange for what they see as a much shorter path to commercial revenue.

Mind Robotics says proceeds will fund: roughly $200 million for M1 pilot manufacturing; $150 million for compute and a dedicated training cluster co-located with Nvidia infrastructure in Texas; the build-out of a combined R&D and pilot-production facility outside Austin; and an aggressive hiring plan targeting 600 employees by the end of 2026.

Market Context: The Industrial Wedge of Physical AI

The global industrial robotics market sits at roughly $55 billion in 2025 and is projected by BCG to surpass $120 billion by 2030. Within that, the segment most analysts are watching — AI-enabled, general-purpose manipulation — is forecast to grow north of 35% annually. The convergence of three forces is driving the surge: a structural US manufacturing labor shortage, reshoring incentives under continuing CHIPS and IRA frameworks, and the maturation of vision-language-action models pioneered by Google's RT-2, Nvidia's GR00T and Physical Intelligence's π0.

Competitive positioning matters here. The humanoid camp — Figure, 1X, Apptronik, Tesla's Optimus — is selling a long-dated vision of universal labor. Legacy players like Boston Dynamics and Agility Robotics own the warehouse logistics niche. Pure-software "robot brain" companies like Skild AI and Physical Intelligence are racing to become the Android of robotics. Mind Robotics is attempting to occupy the middle: a vertically integrated stack — hardware, model, and data flywheel — but with a pragmatic, deployable form factor.

Accel partner Rich Wong framed the thesis bluntly in a LinkedIn post: "Humanoids are a five-year promise. Mobile manipulators are a twelve-month revenue line." Katherine Boyle of a16z's American Dynamism practice added that Scaringe's manufacturing track record is "the single most underpriced asset in the robotics market right now."

Strategic Implications: Rivian as Customer Zero

The most important commercial detail in the announcement is also the simplest: Rivian itself will be Mind Robotics' first customer. The company plans to deploy approximately 500 M1 units across Rivian's Normal, Illinois facility and its under-construction plant in Social Circle, Georgia, beginning in the second half of 2026. That deployment functions as both a flagship case study and a brutally demanding beta site — automotive final assembly is widely considered the hardest test bed in industrial robotics.

If the Rivian rollout works, the doors open quickly. Pilot conversations are reportedly underway with PepsiCo in beverage logistics and Maersk in port operations, and at least one additional automotive OEM is said to be in late-stage commercial discussions. Siemens' strategic investment is the other key enabler: integration with Siemens' PLC and MES ecosystem would let Mind Robotics drop into existing brownfield plants without the painful systems-integration costs that have historically gated industrial robotics adoption.

Longer term, the company is betting on a data flywheel. Robots in the field generate proprietary manipulation data; that data trains the next version of MindOS in a Texas-based cluster; better models drive higher utilization, which in turn justifies more deployments. It is the same playbook Tesla ran for autonomy, applied to factory work.

The Investor Logic

Three things appear to have convinced investors to underwrite a $2 billion valuation on a pre-revenue company. First, execution credibility: whatever one thinks of Rivian's financial trajectory, Scaringe took an EV company from zero to tens of thousands of vehicles annually, which is more than almost any robotics founder can claim. Second, form-factor pragmatism: a wheeled dual-arm system has a far clearer ROI calculation than a $200,000 humanoid that may or may not walk reliably. Third, a real first customer in Rivian, which removes the chicken-and-egg problem that has stalled many robotics startups.

Marc Andreessen captured the mood on X: "If America is going to make things again, we need a new kind of workforce. Mind Robotics is the closest thing I've seen to an answer."

What to Watch Next

The $500 million round positions Mind Robotics as the most credible non-humanoid challenger in physical AI, but the next twelve months will be unforgiving. Investors and would-be customers will be watching four metrics closely: M1 uptime and cycle-time data from the Rivian Normal plant; the launch and developer reception of the Siemens integration SDK; the announcement of a second automotive OEM customer; and any decision to open MindOS to third-party hardware. If 2025 was the year humanoids captured the imagination, 2026 may be the year industrial mobile manipulators capture the purchase orders — and Mind Robotics has just bought itself the runway to find out.