Intelligence Factory

Intelligence Factory (legal name Intelligence Factory Labs, Inc.) is a New York City startup building foundation models for robotic manipulation. It was founded in 2026 and is part of the Y Combinator Spring 2026 (P26) batch.

The company's thesis is that robot hardware is converging and that intelligence is now the main bottleneck for general-purpose robots. It frames the commercial problem as closing the gap between roughly 40% task success rates seen in real-world robot deployments and the 99% reliability it says is needed before a robot has commercial value. Its approach relies on human demonstration data rather than only robot teleoperation data.

Intelligence Factory captures human demonstration data in its own facilities, where people perform tasks while wearing custom gloves that record visual, motor and tactile information (described by the company as capturing what the demonstrators "see, do, and feel"). This human motion is then retargeted to robot motion so that trained manipulation models can be deployed across different robot embodiments, ranging from dexterous multi-finger hands to parallel grippers. The company states that it operates its own data pipeline and is scaling multimodal human demonstration data so that its models generalize across tasks, objects and environments. In July 2026 it published a whitepaper on its website arguing that generalized robotics is fundamentally a data and deployment problem and outlining its "data factory" strategy.

The company's deployments span warehouses, grocery stores and data centers.

The company was co-founded by Yash Sinha (CEO) and Jalaj Shukla (CTO). Sinha studied robotics at ETH Zurich, was previously founder and CTO of Kaio Labs and worked as a Formula 1 simulation engineer. Shukla holds an MS in Robotics and Mechatronics from the University of Pennsylvania, where he was a research assistant in the GRASP Lab working on multimodal manipulation policies; he previously worked on manipulation, VLA models and controls at Dimensional, as a senior robotics engineer (manipulation) at Blue Sky Robotics and as a computer vision engineer at Applied Materials. The company says its team draws on people from ETH Zurich, the University of Pennsylvania, Harvard University, NVIDIA, Formula 1 and Ansys.

The team has 10 people, with open roles in New York (founding manipulation engineer, GTM/operations intern) and India (data engineer, mechanical engineer). The company has access to about USD 750K in cloud compute credits across Azure, AWS and GCP plus NVIDIA Inception program benefits.