SoftBank and Nvidia consider $1B investment in Skild AI to build a universal “brain” for robots
The funding push signals growing confidence in foundation models for robotics, a shift that could make machines cheaper, more flexible, and easier to deploy.
Robots are everywhere in factories and warehouses, but most still need custom software and hardware to do even simple tasks, which makes them expensive and hard to scale.
Now, Skild AI wants to fix that by building a universal “brain” for robots, basically software that can work across many types of machines. That idea is ambitious enough that SoftBank and Nvidia are reportedly preparing to back it with more than $1 billion.
According to Reuters, the new funding round, if it comes through, would value Skild AI at about $14 billion, nearly triple its $4.7 billion valuation earlier this year. But the company isn’t building robot hardware. Instead, it’s developing a “robot-agnostic foundation model,” Skild Brain, that can be plugged into different machines to give them advanced behaviors.
Louis Eriakha
In July 2025, Skild showed robots using its model to pick up objects, navigate cluttered environments, climb stairs, and stay balanced. The goal is to let manufacturers license this software instead of writing custom code for every machine, making robots easier to deploy across industries.
SoftBank’s interest also aligns with its growing robotics push, including its recent $5.4 billion deal to buy ABB’s robotics business. Nvidia’s involvement strengthens its footprint in real-world AI, since models like Skild Brain depend heavily on high-performance chips.
If this approach works, the robotics industry could shift from custom-built systems to machines running shared, general-purpose intelligence, a change that lowers costs and accelerates adoption.
Skild AI’s potential $14 billion backing isn’t just another funding milestone. It signals that investors are pushing from speculative robotics hype toward serious bets on general-purpose robot intelligence.
Kelechi Edeh
