Problem

Forge Robotics addresses the structural labor and flexibility problem in metal manufacturing. Its current materials point to roughly 415,000 unfilled U.S. manufacturing jobs today and an estimated 3.8 million additional workers needed by 2033; welding is especially constrained, with the company citing approximately 400,000 unfilled welding positions expected to rise to 600,000 by 2030. At the same time, manufacturers are trying to reshore production, while conventional industrial robots remain limited to tightly defined, human-programmed tasks. Variation in part geometry can leave roughly 10% of welded parts needing rework, undermining the economics of both automation and domestic production.

The killer use case is high-mix metal fabrication: taking raw or cut metal stock and turning it into finished welded parts and assemblies without depending on scarce skilled welders or repeatedly reprogramming a robot for every variation. This is particularly relevant to manufacturers producing defense, construction, automotive, and other industrial components where part volumes, shapes, and jobs change frequently.

Product / Service

Forge’s initial product is a robotics intelligence layer for welding. A robot-mounted vision system identifies features and weld seams, creates a sub-millimeter three-dimensional map of the workpiece, and continuously directs the welding arm’s motion in real time. That changes a conventional robot from a rigid machine following a taught path into an adaptive system that can see and respond to the actual part.

The broader delivery model is a modular autonomous factory: shipping-container-sized, self-contained production modules that can be configured in software and added together to expand capacity without proportionally expanding headcount. Modules can handle welding, machining, fastening, finishing, scanning, and logistics; they plan, execute, inspect, and route parts autonomously. Forge says modules can be operational within a week, run continuously without operators, and achieve 99% first-pass yield through in-line scanning and verification, promising faster deployment, less manual programming, fewer defects, and more scalable domestic production.

Market

Forge competes in industrial robotics and robotic process automation, with its initial beachhead in AI-enabled robotic welding and metal fabrication. Its competitive set includes direct adaptive-welding providers such as Path Robotics, easier-to-program cobot-welding offerings such as Hirebotics, and established automation and sensing suppliers such as Lincoln Electric and SERVO-ROBOT. Forge’s intended differentiation is the combination of real-time vision and control with a broader modular factory architecture designed for changing parts and production steps, rather than a fixed, manually programmed cell.

The company was founded in 2025, is an active Fall 2025 Y Combinator company, and is publicly presenting a roadmap from welding intelligence to fully autonomous factory units. Its YC profile says it is taking as many pilot customers as possible, so the available evidence indicates an early pilot or pre-scale company rather than one with demonstrated commercial scale. Public materials reviewed do not disclose revenue, customer counts, or completed deployments, so Forge should be treated as likely pre-revenue or at least commercially unproven as of August 1, 2026, despite its Y Combinator backing and active customer-development effort.