At 1872, the founders are applying lessons from that manufacturing environment to steel components that require less precision than aerospace hardware. The company aims to automate most of its fabrication process by 2027, but Summers said eliminating human involvement entirely is not the objective. “We’re building towards autonomy, but we’re not necessarily building in a dogmatic fashion towards full autonomy,” Summers told Ars. “We may achieve 80 percent autonomous operations, and we may decide that it makes sense to stop there because there’s just diminishing returns to go to full 100 percent.”
Steel skids provide a relatively forgiving starting point. Summers described them as lower-precision components compared with aerospace parts, allowing the company to develop its automation systems around products that can tolerate more variation while still serving infrastructure projects.
The approach also targets a labor constraint in welding and fabrication. The American Welding Society estimates the U.S. will need 320,500 new welding professionals by 2029 as workers retire and demand increases across industries including data centers, semiconductor manufacturing and shipbuilding. “The problem that we are trying to solve is how do we build more things with a decreasing pool of skilled labor to do it with, and this is a good place to start in terms of implementing automation,” Summers said.
1872 began by using human workers to establish its skid production process before introducing automation. It is working with Columbus-based Path Robotics on robotic arms that perform arc welding, one of several physical manufacturing tasks the startup wants to incorporate into a more automated factory.
Path Robotics says its systems achieve first-pass yields between 95% and 100%, meaning most parts clear inspection without needing rework or becoming scrap. The company also reports that its robots spend about 70% of their working time actively welding, compared with 10% to 12% for human workers who must also position materials and adjust equipment.
Path Robotics estimates its automated systems can weld for about $0.12 per inch, compared with roughly $0.78 for manual welding, representing an 85% reduction in welding costs.
But welding itself is only one part of the production bottleneck 1872 wants to address. Summers said welding a skid requires roughly two to four hours, while preparing and assembling the components beforehand can take four to five days. Automating more of that preparation could therefore have a larger effect on overall production time. “The whole name of the game is how do we keep that machine fed,” Summers said.
The software layer is central to how 1872 plans to coordinate those operations. Its proposed system begins with Architect, AI-driven software designed to take a customer’s digital design files and turn them into a manufacturing plan covering areas including material sourcing, pricing and scheduling.
A second system called Conductor is intended to carry out that plan on the factory floor by coordinating equipment and material movement. The setup could eventually involve robotic arms, autonomous vehicles moving components between workstations and robots traveling along production lines. “Our differentiator is our ability to take robotic systems that are either built by us or built by partners like Path Robotics, pull them into a single system and orchestrate them together in a way in which you can really achieve a seamless operation,” Summers said.
1872 is using customized versions of existing AI models, including large language models, and plans to add physics-based models representing manufacturing processes. Production runs are also expected to generate additional data that can be used to improve the AI systems over time.
The founders trace part of that approach to their work at SpaceX. Summers said software engineers worked alongside hardware engineers on Raptor not only to develop the engine but also to create the systems needed to manufacture it while designs were changing rapidly. According to Summers, the Raptor team moved from its first heavily instrumented full-scale concept to a production engine within three years.
1872 now wants to bring a similar combination of software and physical production to steel fabrication. Summers said the $15 million seed round is intended to support Factory One, development of the automation hardware and software, and the company’s effort to reach profitability.
The company could eventually move beyond skids into products including structural frames and enclosures. Before that expansion, its immediate work is focused on replacing portions of the manual process at Factory One with automation. “Over the next 12 months is when we expect to start to really peel back the layers of the onion on the manual processing, and implement automations not only on the physical production floor but digitally as well,” Summers said.
For 1872, that means proving that AI software and robotics can coordinate enough of a real manufacturing operation to increase output without requiring full autonomy. The first test is intentionally straightforward: make steel skids more efficiently, then determine how much further the same system can extend across other components.
This analysis is based on reporting from arsTECHNICA.
Image courtesy of The Business Journals.
This article was generated with AI assistance and reviewed for accuracy and quality.
