Skild AI is building robot foundation models into general-purpose control
A July 9, 2024 Series A report and general-purpose robot brain framing give Skild AI a foundation-model robotics anchor.
Forbes reported on July 9, 2024 that Skild AI raised a $300 million Series A to build a general-purpose robot brain. The reporting placed the company valuation at $1.5 billion, giving Skild a large capital base before public customer deployments were fully visible.
Skild was founded in 2023 by Deepak Pathak and Abhinav Gupta, former Carnegie Mellon University robotics professors. That origin gives the company serious robotics-learning credentials, but the commercial question is whether software can transfer across useful robot bodies and customer tasks.
The company is trying to sell a hardware-agnostic intelligence layer: software that can help different machines perceive, plan, and act without each robot requiring a separate control stack. That is powerful if it works, but difficult to validate from funding and model demos alone.
The competitive field includes Physical Intelligence, Covariant?s RFM work under Amazon, Google DeepMind robotics, NVIDIA Isaac and GR00T, Figure?s Helix-style stack, Genesis AI, Sereact, and internal foundation-model programs inside large robotics companies. Skild?s distinction is the explicit general-purpose brain framing backed by major capital and academic robotics founders.
Public material does not show customer deployment count, task success rate by robot body, autonomy intervention frequency, field uptime, paid customer retention, pricing, renewal rate, deployment density, or training-data composition. The round is a strong strategic signal, not proof of operating scale.
Skild?s strategic test is whether robot intelligence becomes a bought software layer instead of a custom stack inside each machine. If the company can show task transfer across real customer robots, Skild becomes infrastructure for physical AI rather than another foundation-model promise.
Have a robotics update Korthos should review? Send news, deployments, product releases, funding rounds, research, or media to tips@korthos.xyz or reach out on X at @agkorthos.