Gritt Launches with $32.4M to Bring Physical AI to Infrastructure Construction

Gritt has emerged from stealth with $32.4 million in funding to develop AI-powered robotic systems capable of performing repetitive and physically demanding work on large construction sites.

The financing includes a $26 million Series A led by Obvious Ventures, with participation from Union Square Ventures and Active Impact Investments. Earlier investors First Round Capital, Climactic, Congruent Ventures, and VSC Ventures also supported the company.

Gritt is initially deploying its technology across large-scale solar projects, where its machines are already being used to move, position, and assemble construction materials. The company plans to expand into data center construction and other major infrastructure projects.

Building Robots for Unpredictable Environments

Industrial automation has traditionally worked best inside controlled environments such as factories and warehouses. Construction presents a considerably harder robotics problem.

Outdoor jobsites change continuously. Machines must operate across mud, snow, dust, uneven terrain, variable lighting, and shifting layouts. Materials, equipment, and workers may also move throughout the day, making it difficult for a robot to depend on a fixed map or tightly choreographed workflow.

Gritt is developing what it describes as a shared AI backbone for construction work. The system combines robotic manipulation, spatial perception, site data, and decision-making software so machines can perform tasks in environments that cannot be fully predicted in advance.

Rather than manufacturing an entirely new fleet of construction vehicles, Gritt’s robotic arms and AI systems are designed to connect with equipment contractors already use, including skid steers and forklifts. Off-the-shelf attachments and limited custom hardware allow the system to be introduced without requiring construction companies to replace familiar machinery.

This approach could also simplify maintenance and deployment, since crews can continue relying on established equipment platforms while adding autonomous capabilities for selected tasks.

From Material Handling to Assembly

Gritt’s machines are being developed to perform several categories of construction work through the same underlying system.

Its physical AI capabilities include picking up and positioning components, transporting materials around a site, and assembling parts into finished structures. The company is also developing spatial AI functions for inspecting completed work, maintaining project records, monitoring site conditions, and recommending what should happen next.

The system is designed to place and manipulate materials with millimeter-level precision, even as conditions around the machine change.

According to Gritt, its machines currently collect terabytes of spatial and operational data each day. That information can be used to improve perception and task execution while creating a clearer record of material movement, completed work, and changing site conditions.

This data layer may eventually become as important as the robotic hardware. Construction supervisors frequently make decisions using incomplete or delayed information. A system that continuously observes a site could help identify bottlenecks, verify progress, and flag emerging problems before they disrupt a project.

Starting with Large-Scale Solar Construction

Gritt is beginning with utility-scale solar, a sector that requires crews to repeat precise physical movements across enormous outdoor sites.

The company says its systems have autonomously placed tens of thousands of solar panels without breaking a panel. Its machines are operating across more than seven jobsites and have achieved efficiency gains of up to four times without requiring additional labor.

Solar construction offers a useful testing ground for physical AI because the work combines repetition with difficult environmental conditions. Components must be transported and positioned accurately, but the machines performing those tasks may be operating across hundreds of acres of uneven terrain.

Gritt says its equipment is already being used by major U.S. engineering, procurement, and construction companies. The company is also developing capabilities for concrete and rebar work, with data centers and broader infrastructure construction identified as future markets.

Addressing Construction’s Labor and Productivity Challenges

The funding arrives as construction companies face a widening gap between infrastructure demand and workforce availability.

More than 41% of the U.S. construction workforce is expected to retire by 2031, according to figures cited by Gritt. At the same time, major spending is planned for energy infrastructure, transportation, data centers, manufacturing facilities, and other large projects.

Construction productivity has also struggled to keep pace with other industries. Although software has improved project planning and administration, much of the work performed on active jobsites still depends on people manually moving, positioning, and assembling materials.

Gritt is targeting the repetitive, high-volume tasks that remain difficult to automate but can consume substantial amounts of time and physical effort.

The company is positioning its technology as a way to supplement construction crews rather than replace entire jobsites with autonomous machines. Workers could supervise robotic systems and focus on tasks requiring judgment, coordination, and specialized skills while machines perform thousands of repetitive movements.

Creating an Intelligence Layer for the Jobsite

Gritt’s longer-term goal extends beyond automating individual construction activities.

As its machines operate, they gather information about site layouts, equipment movements, installed materials, and environmental conditions. The company plans to use this data to build a decision-making system that can understand how work is progressing across an entire project.

Such a system could eventually help supervisors anticipate material shortages, detect scheduling conflicts, recommend the next sequence of work, and coordinate resources across multiple crews and machines.

This would move Gritt closer to an agentic construction platform in which AI does not simply execute commands but continuously evaluates conditions and helps determine how a project should proceed.

Each deployment may also provide training data for future sites. Gritt says tasks that initially required months of development can now be adapted within days as the system accumulates experience.

Future Implications for Construction

Gritt’s funding reflects growing interest in physical AI systems that can operate beyond controlled factories and warehouses.

As these technologies mature, construction firms could automate more repetitive material-handling and assembly tasks while giving supervisors better real-time visibility into project progress, equipment use, and site conditions. This may help reduce delays, improve safety, and allow limited skilled labor to focus on work requiring judgment and coordination.

The broader opportunity is an increasingly adaptive jobsite where machines learn across deployments and help infrastructure projects move from planning to completion more efficiently.