Skild AI

Skild AI was founded in 2023 by former Carnegie Mellon professors Deepak Pathak and Abhinav Gupta to build a general-purpose robotics brain. The thesis is that physical AI should be omni-bodied - a single model that controls different robot morphologies and generalizes across tasks - trained on orders of magnitude more robot data than competing efforts. Skild's product suite today covers security and inspection robots in unstructured environments, mobile manipulation, and autonomous packing, all driven by Skild Brain. In July 2024 Skild closed a $300M Series A led by Lightspeed, Coatue, SoftBank, and Jeff Bezos at a $1.5B valuation, and in April 2025 added another roughly $500M round led by SoftBank, pushing total reported funding past $1.4B. The company has offices in Pittsburgh, the Bay Area, and Bengaluru.

With Physical Intelligence, Skild is one of the two most heavily funded 'robotics brain' pure-plays - if one of them wins broad adoption, the humanoid market collapses into a model-vs-hardware split similar to Android/OEMs.

skild.ai Robotics Foundation Models

United Statesprivate

At a glance

Total raised

$2.8B

last round series c

Headcount

51-200

approximate · public range

Headquarters

Pittsburgh

United States

Founded

2023

3 years active

Status

Private

Robotics Foundation Models

Latest valuation

$14B

Series C · 2026

For subscribers

Subscribe to see the full profile

Including Funding rounds, Bull / Bear thesis, Stock + earnings, Roster changes, Patents, News, and Open roles.

Founding

50% off

/month

94 of 100 spots left

Full

$10

/month

Available once all 100 Founding Member spots are claimed.

Get full access

Already subscribed? Sign in →

Funding history

Total raised

$2.8B

Rounds tracked

Latest stage

Series C

Latest valuation

$14B

Date	Stage	Amount	Valuation	Lead investors
Apr 1, 2025	Series B	$500M	$4.5B	SoftBank Group

Founder backgrounds

Abhinav Gupta

Co-founder & President

A professor in Carnegie Mellon University's Robotics Institute, Gupta's academic research has centered on learning generalizable robotic behaviors from large-scale, unstructured data — particularly internet video. Before co-founding Skild AI, he was a research scientist at Facebook AI Research (FAIR) and held a postdoctoral position at UC Berkeley after earning his PhD from the University of Maryland. His prior work on learning manipulation skills from in-the-wild video directly informs the Skild Brain approach: a single foundation model meant to generalize across robots, tasks, and environments rather than being fine-tuned per deployment.

Deepak Pathak

Co-founder & CEO

Before founding Skild AI, Deepak Pathak was an Assistant Professor at Carnegie Mellon's Robotics Institute, where his lab directly developed the generalist-robot research that became the company's foundation. He earned his PhD at UC Berkeley under Trevor Darrell and Jitendra Malik, producing influential work on self-supervised visual learning (including the highly cited "Learning to See by Moving") and reinforcement learning for embodied agents. He completed his undergraduate degree at IIT Kanpur. Pathak's academic posture — publishing open-source robotics research while simultaneously building toward a commercial foundation model — reflects a tight coupling between scientific rigor and product ambition.

AI bull / bear

Bull case

Skild AI has assembled the capital, talent, and strategic partnerships to become the Android of robotics. With over $1.4B raised and a valuation now exceeding $14B after the recent Series C led by SoftBank and NVIDIA, the company has unprecedented firepower to execute on its omni-bodied foundation model vision. The acquisition of Zebra's Fetch Robotics division adds immediate commercialization muscle—fleet management software, customer relationships, and deployed hardware—turning Skild from a pure research play into an integrated robotics stack provider. NVIDIA's co-lead in the Series C is particularly significant: it signals deep technical validation and potential silicon/toolchain co-optimization. If Skild's model-agnostic brain proves it can generalize across morphologies—from AMRs to humanoids—it captures value across the entire physical AI stack while hardware OEMs commoditize. The warehouse automation wedge offers a massive, near-term TAM with clear ROI, funding the longer roadmap to general-purpose manipulation.

Bear case

Skild's $14B+ valuation assumes winner-take-most dynamics in robotics foundation models, but the omni-bodied thesis remains unproven at scale. Physical Intelligence raised similar capital and is pursuing the same universal brain narrative—splitting mindshare, talent, and potential customers in a winner-take-most market creates existential risk for both. The Fetch acquisition may distract from core R&D; integrating legacy fleet software and supporting legacy hardware customers pulls resources toward sustaining engineering rather than frontier model development. SoftBank's lead also raises execution risk—the firm's track record in robotics (Boston Dynamics sale, Pepper writedown) suggests appetite for hype over sustainable unit economics. Warehouse automation is commoditizing fast, and customers like Amazon build in-house. If the foundation model doesn't generalize beyond structured environments within 18–24 months, Skild burns through capital competing with vertical-specific incumbents that already have deployment scale and cheaper, task-specific models.

Generated by claude-sonnet-4-5-20250929 · April 25, 2026Refreshes quarterly · see disclosures

In the news

Skild AI · last 30 days

No articles ingested yet for Skild AI. Once the hourly news pipeline is live, every article the classifier tags as mentioning this company appears here with its one-line AI summary and sentiment.

Leadership

Deepak Pathak

Co-founder & CEO

in 𝕏

Abhinav Gupta

Co-founder & President

—

Products

Skild Brain foundation model for robots
Omni-bodied robotics foundation model that controls any robot without prior knowledge of its exact body form, including quadrupeds, humanoids, tabletop arms, and mobile manipulators. The architecture combines a high-level decision-maker with a low-level controller that translates commands into precise joint angles and motor torques. Pre-trained on physics-based synthetic data generation and human videos from the internet—sources almost infinitely more scalable than real-world teleoperation—then post-trained on targeted real-world data. First deployed on NVIDIA's Blackwell GPU production lines with Foxconn in 2026, controlling dual robotic arms for complex assembly operations.
Security / inspection robot platform
A quadruped robot platform for security and inspection, powered by the Skild Brain—an omni-bodied foundation model that enables robots to navigate unstructured, dangerous environments. The platform includes specialized software modules for the security and inspection vertical, alongside licensing through cloud-based APIs where customers pay for robot intelligence rather than developing proprietary control systems. Skild's approach enables successful robot deployment on low-cost systems costing $4,000–$15,000, versus conventionally customized systems costing $250,000+.
Mobile manipulation platform
Mobile robot platform powered by Skild Brain that enables development of AI algorithms and applications for robotic manipulation via API. The AI can execute low-level skills like grasping, handover, and navigation on mobile platforms. The foundation model works on quadrupeds, humanoids, table-top, mobile manipulators and more, enabling developers to create robotics applications using simple API calls, facilitating broader access to robotics development. LG CNS, the IT arm of South Korea's LG Group, has partnered with Skild for industrial applications.
Autonomous packing system
Warehouse-scale robotic packing system designed to automate the repetitive work of packing and sorting items using nimble robotic skills. Built on the Skild Brain, the industry's first omnibodied AI software designed to operate without prior knowledge of a robot's exact body form, enabling packing coordinated through a unified orchestration layer alongside picking and material movement. Following Skild AI's acquisition of Zebra Technologies' Robotics Automation business and its Symmetry Fulfillment orchestration platform, the system now has infrastructure to deploy and orchestrate entire fleets of Skild Brain-powered robots at enterprise scale.

Open roles

View all careers ↗