The Future of Steel Fabrication: Why AI-Native Companies Will Win the Next Decade

Two decades ago, a small online bookseller called Amazon told traditional bookstores what was about to happen. Most of those stores tried to respond by adding an online order form to a business that was still organized around physical shelves, paper inventory, and store managers. They did not lose because they ignored the internet. They lost because the company that won was built around the internet from the first day. It was internet-native, and everyone else was a legacy shop with a website.

Steel fabrication is approaching the same kind of inflection point. The question is not whether AI will be useful in our industry, because it already is, in the same surface-level way that the internet was useful to bookstores in 2002. The real question is which fabricators will rebuild the work itself around AI, and which will simply add AI as a feature on top of the same processes they have always run.

This article shares American Katerra’s view, from inside a working steel fabricator, of what the next decade actually looks like, and why we believe the companies that come out on top will be the ones that are AI-native from day one.

There are two ways to bring AI into a fabrication business. From the outside they look similar. They lead to very different outcomes.

A bolted-on AI shop keeps the existing org chart, the existing handoffs, and the existing pace, and then hands each person an AI tool. The estimator gets a copilot. The detailer gets a smarter clash-detection plug-in. The project manager gets an AI summary of the meeting. Each person individually gets faster. The overall workflow, the contracts, the dependencies between trades, and the way decisions move through the company all stay the same. The fabricator looks modern, but the underlying production system is still the one from 1995.

An AI-native fabricator does something different. It treats AI the way Amazon treated software in 2000, as the load-bearing structure of the operation rather than as decoration. Estimating, detailing, scheduling, RFI handling, supplier coordination, lead-time tracking, and quality control are designed so that AI does the first pass on every routine document, every model check, and every status update. Human time is reserved for the work AI cannot do: judgment calls, gray-area decisions, and final accountability.

Both fabricators will say they “use AI” on their website. Only one of them is built around it. On a long enough horizon, the second one wins, for the same reason Amazon did.

Here is how we expect the next decade to play out across the steel fabrication value chain. Some of this is already in motion. Some of it is two or three years away.

• Estimating and takeoff. This is the easiest shift to predict because it is already starting. AI-assisted estimating tools pull quantities and structural assumptions from drawings and BIM models in minutes rather than days. American Katerra already uses BIM-based estimating with tools such as Trimble EC-CAD, and the AI layer will push that workflow further. The estimator’s role moves from re-entering numbers to challenging the assumptions behind them and pricing risk. Faster bids, more bids reviewed, fewer mistakes carried into contracts.
• Structural steel detailing. This is the function most affected. Today, detailing is the critical path of almost every steel project. In an AI-native shop, AI handles the routine geometric and rules-based decisions inside tools like Tekla and Revit, and surfaces only the items that need a qualified detailer’s judgment. The detailer’s time concentrates on connections, code interpretation, and coordination, which are the parts that actually require experience.
• Shop drawings, RFIs, and submittals. AI drafts the documents, and the project engineer reviews and signs. The engineer’s day shifts away from formatting and chasing wording, and toward solving the problem behind each RFI. The cycle time on every submittal package starts to compress.
• Production scheduling and shop-floor QC. AI watches the schedule continuously, flags bottlenecks before they cascade, and reconciles what the shop is actually producing against the plan. QC inspections become AI-assisted but still human-signed. The inspector still owns the call.
• Supply chain, lead time, and material visibility. This is where the buyer feels the change first. Owners and general contractors do not search for “AI in steel fabrication.” They search for cost estimation, lead time, and supply chain reliability. An AI-native fabricator can give a clearer commitment on lead time because it has continuous visibility across mill, processor, and shop. That is the language a general contractor actually buys in.

Our view: AI does not lower the bar for the people in a steel shop. It raises it. The routine work disappears, and what is left is the work that requires real experience, real judgment, and real accountability.

Two kinds of decisions will continue to belong to people, no matter how capable the tools become.

The first is the gray-area call. Every experienced fabricator knows the situation. One interpretation is technically fine, another is technically a code problem, and the right answer depends on context that is not written down anywhere. AI can summarize the relevant code section and surface precedent, but it should not be the one signing the decision. The qualified detailer, the licensed engineer, and the project manager who has to face the client on Monday morning are the ones who make that call, and that does not change.

The second is the final commitment. When a general contractor asks a fabricator to commit to a price and a lead time, they are buying a promise. AI can prepare every input that goes into that promise, including pricing, schedule, material availability, and capacity. The promise itself is still made by a person, because only a person can be held accountable for it. AI-native does not mean unaccountable. If anything, it makes the accountability cleaner. The person making the call has better information and fewer excuses.

In structural steel, it has historically taken roughly ten years to develop a fully capable detailer, fabricator, or project engineer. The knowledge is mostly tacit. You learn by watching, by being corrected, and by being on the shop floor when things go wrong. That model worked when the industry could afford to wait ten years. It will not work for the volume of work coming into the US market over the next decade.

This is one of the areas where AI changes the curve most dramatically. When AI can pull the right code section, summarize the relevant precedent from your own past projects, and explain why a particular connection detail was chosen on a similar job last year, the new hire is not waiting for a senior person to be available. The senior person is still the one validating the answer. But the apprentice gets to the validation stage in three years instead of ten. The human judgment layer at the top still has to be developed the hard way. The information layer underneath it does not.

We believe this is how the structural steel workforce gap actually gets solved in the United States. Not by recruiting ten times as many senior detailers, but by building shops where less experienced people, supported by AI and supervised by senior people, can produce work at a senior level much earlier in their careers.

Our parent company has a phrase that guides how we think about this: “Beyond the Information Revolution. Happiness for Everyone through the Innovation of Products.” Most of what gets called “AI” today is still part of the information revolution that started in the 1990s. Faster search, faster summarization, faster answers. That is necessary work, but it is not the interesting part for our industry.

The interesting part is what happens when that information layer is wired directly into how things are physically made. How steel is detailed, scheduled, cut, welded, shipped from our San Antonio operation, and erected on a job site in Texas. That is the shift American Katerra is trying to be early on. We are not aiming to be a steel fabricator with an AI feature page on our website. We want to be a steel fabricator whose entire operation is designed so that AI does the first draft of everything, and our people do the work only humans can do: making the call, taking responsibility, and standing behind the steel that leaves our shop.

Two principles guide how we are building the company.

First, we lead with the physical product. The reason a project owner hires American Katerra is to get structural steel that is detailed correctly, fabricated to AISC standards, and delivered when we said we would. AI does not change that. It is the means, not the message. Our customers buy steel. How we make it well is our problem, not theirs.

Second, AI lives inside the operation, not bolted to it. Estimating, detailing, scheduling, document handling, and supplier coordination are all designed so that AI does the first pass. We hire and train for people who are comfortable directing AI tools rather than competing with them, and we hold them to the same accountability standard as any senior person in a traditional shop.

Under the leadership of President Toyokazu Yamaguchi, and backed by the manufacturing depth of our parent company Yamaguchi Heavy Industries in Japan, this is the operating model we are taking out into the US market. It is also the model that supports our long-term direction, which includes our intention to list on the Texas Stock Exchange (TXSE) as the company matures.

The next decade of structural steel will not be decided by who has the most impressive AI demo on their homepage. It will be decided by which shops actually delivered, on time and on spec, on the projects that mattered. We are building American Katerra so that, when that question is asked five and ten years from now, the answer is on the right side of the line.

Planning a U.S. project that needs a structural steel partner with deep Japanese manufacturing roots and on-the-ground execution in Texas? Request our Capability Statement for an overview of our fabrication capabilities, project experience, and quality processes.