SkildAI Raises $1.4B for Robotic Foundation Models — Physical AI Gets Its Mega-Round
SoftBank and Nvidia co-lead a $1.4B Series C at $14B valuation for robotic AI foundation models. Physical AI is now attracting language-model-scale capital.
The Round
SkildAI, which builds AI foundation models to power robots, has raised $1.4 billion in a Series C at a $14 billion valuation. The round was co-led by SoftBank and Nvidia.
This is among the largest robotic AI funding rounds on record.
What SkildAI Does
While most frontier AI labs focus on language, code, and reasoning, SkildAI builds foundation models for physical interaction — models that allow robots to:
- Manipulate objects in unstructured environments
- Navigate complex physical spaces
- Learn generalizable motor skills that transfer across different robot types
- Adapt to new tasks without complete retraining
Think of it as GPT, but for robot hands and legs instead of text.
Context: 2026 AI Funding Scale
| Company | Round | Amount | Valuation |
|---|---|---|---|
| Anthropic | Series G | $30B | $380B |
| OpenAI | Closing | $100B | $850B |
| SkildAI | Series C | $1.4B | $14B |
| ElevenLabs | Series D | $500M | $11B |
17 US AI companies have raised $100M+ in just the first two months of 2026.
Why SoftBank and Nvidia Co-Led
SoftBank has bet heavily on robotics through its Vision Fund and its ownership of Boston Dynamics. A foundation model that makes diverse robots more capable amplifies the value of every robotics investment in SoftBank’s portfolio.
Nvidia recently launched Nemotron 3 (open models for agentic AI) and Isaac (robotics simulation platform). Investing in SkildAI extends Nvidia’s reach from silicon to the software layer that makes robots intelligent — completing a vertical stack from chips to robot brains.
The Physical AI Thesis
The investment thesis: language models transformed how we interact with information. Physical AI foundation models will transform how machines interact with the physical world.
The market is enormous:
- Manufacturing — flexible robot workers that adapt to new products
- Logistics — warehouse and delivery robots
- Healthcare — surgical assistance and elder care
- Agriculture — autonomous harvesting and monitoring
- Construction — dangerous or repetitive physical tasks
What’s Different From Previous Robotics Hype
Previous robotics waves failed because each robot needed custom programming for each task. Foundation models change this by enabling:
- Transfer learning — skills learned in simulation apply to real-world robots
- Generalization — one model powers many different robot types
- Natural language control — tell the robot what to do, don’t program every movement
This is the same paradigm shift that made LLMs transformative: from narrow, task-specific systems to general-purpose capabilities.
What to Watch
- SkildAI’s first production deployments and which industries they target
- Whether the model generalizes across robot hardware from different manufacturers
- Competition from Google DeepMind (RT-X), Tesla (Optimus), and Figure AI
- How Nvidia integrates SkildAI capabilities with its Isaac robotics platform
Physical AI just got its ChatGPT moment in funding. The products are next.
Sources: TechCrunch, ContentGrip