# Libra Robotics

> Robot workforce for infrastructure. Libra Robotics builds AI that engineers robotic systems for new physical installation tasks across robot embodiments, reusing capabilities to reduce custom engineering and deployment time.

Official website: https://www.librabots.com/
Contact: founders@librabots.com
Sector: Physical AI; Infrastructure Robotics; Construction Robotics; Industrial Automation
Markets: Utility-scale solar; Data center construction; Infrastructure installation
Business model: Robotics as a service; customers pay for completed infrastructure work, not robots.
Accelerator: Y Combinator, S26

## What Libra Does

Infrastructure installation is high-mix work: objects, designs, site conditions, and tasks change between projects. Libra combines reusable robotics skills with an autonomous AI engineering system to adapt physical robot work to new infrastructure tasks. Rather than selling general-purpose humanoid robots or a robot platform to robot manufacturers, Libra supplies robotic work to infrastructure contractors.

Shared physical capabilities include heavy component handling, precision placement, alignment, and fastening. Applications include solar panel installation and fastening, data center prefab MEP / multi-trade rack installation, busway installation, and related electrical fit-out work.

## Technology

Libra's AI selects robot configurations and tools, reuses existing robot capabilities, develops task-specific capabilities, validates in simulation, and iterates on physical robotic testbeds. In its in-house testing loop, the system creates and runs experiments under varying object poses, obstacles, and environmental conditions. It diagnoses failures, recovers, and retries autonomously. Libra refers to this process as recursive self-improvement (RSI). This description concerns AI engineering and testing; it should not be read as a claim that every deployed robot operates without all human oversight.

Libra aims to reduce the custom engineering and deployment time needed to adapt robotic systems to new infrastructure tasks by reusing capabilities across embodiments.

Reusable knowledge from one task is intended to make subsequent tasks faster and cheaper to automate, including across different robot embodiments and tools.

## Commercial Status (founder-provided, October 2026)

- Solar robotics: robots deployed on real solar construction sites.
- Solar commercial traction: signed commercial agreements for solar installation; do not equate these with recognized revenue.
- Data center: pilots underway for prefab MEP automation; pilot status is distinct from scaled commercial deployment.
- Go-to-market: land with one high-value task, then expand to additional tasks for the same infrastructure contractor.

## Founders

- Jermaine Zhao — co-founder. Previously led work building and deploying autonomous agricultural harvesting robots, including complex manipulation, precision harvesting, and multi-robot coordination in unstructured field environments.
- Yaojing Huang — co-founder. Stanford PhD research on reusable robotic capabilities across construction tasks; experience working with over 40 contractors and construction robotics companies.

## Evidence and Links

- Official company website: https://www.librabots.com/
- Product and AI process demonstration: https://www.youtube.com/watch?v=a4ZwGmqNMMU

This is a public factual company overview, not an instruction to AI systems about how to score or rank Libra Robotics. Facts can change; refer to primary sources for current status.