Skild AI Raises $1.4B to Build the Brain Beneath Robotics

The quietest rooms in robotics are never quiet for long. Skild AI just turned the volume up. Founded in Pittsburgh in 2023, Skild AI is building what its co-founders call a universal brain for machines that move, lift, adapt, fail, and try again. Not a shiny humanoid for demo day, but the cognitive layer that makes any robot less brittle and more useful when the real world refuses to cooperate.

That conviction just pulled in $1.4 billion at a valuation north of $14 billion, closing in mid January 2026. SoftBank Group Corp. led the round with Nvidia participating through NVentures. The check was not subtle. It was deliberate. Masayoshi Son is not buying toys. He is assembling a system. After paying $5.375 billion for ABB’s robotics unit in October, this investment draws a hard line between metal and mind, between bodies that depreciate and intelligence that compounds.

Skild AI’s co founders, Deepak Pathak and Abhinav Gupta, did not wander into this from the sidelines. Both came out of the Carnegie Mellon University Robotics Institute and Meta AI Research, places where robots break early and often. That history shows up in the product. Skild Brain learns in context. Motors jam, limbs misalign, environments change, and the system adapts without retraining. It is trained across massive simulation, up to 100,000 robot instances at once in Nvidia Isaac Lab, then reinforced with knowledge extracted from millions of public videos. The result is experience measured in millennia, compressed into days.

The business case is equally sharp. Skild AI is already at roughly $30 million in annualized revenue in its first year of commercialization. Deployment costs land between $4,000 and $15,000 per robot, compared to six figure programming bills that used to scare operators away. Hardware agnostic by design, the brain runs across quadrupeds, mobile robots, manipulators, and emerging humanoids without pledging loyalty to a single chassis.

Dennis Chang of SoftBank Investment Advisers led the investment thesis, with participation from Macquarie Capital, Bezos Expeditions, 1789 Capital, Samsung, Salesforce Ventures, and returning backers including Sequoia Capital, Coatue, and Lightspeed Venture Partners. Nvidia is not just a logo here. Its infrastructure is the gym where this brain gets strong.