Classification society ABS and Radian Forge, a startup based in Portsmouth, Virginia, have started a joint development project to test a question that matters to anyone who has tried to get a printed metal part accepted for real service: can the data a machine records while it builds a part count as evidence that the part is good? The ABS announcement, dated November 12, 2026 from Houston, says the project will look at whether digital manufacturing data can be used to qualify wire arc additive manufacturing (WAAM) metal components for marine and offshore use.

That sounds like a small procedural change. It isn't. Qualification is where many additive projects stall, and this project tests whether the evidence a qualification rests on could come from the build itself instead of only from what happens to the part afterward.

What the Project Actually Examines

According to ABS, the JDP will examine whether four kinds of information give a reliable basis for verifying and validating WAAM parts:

  • Process parameters: the settings the machine ran with while depositing metal.
  • Real-time sensor data: what monitoring equipment recorded during the build.
  • Inspection records: the documented results of checking the finished component.
  • Material testing results: the measured properties of the deposited metal.

The VoxelMatters report on the project, published November 14, describes the same scope: these data streams taken together as a possible basis for verification. The important word in both accounts is "whether." This is an evaluation, not a new rule. Neither ABS nor Radian Forge says that sensor logs will replace physical testing. The project is meant to find out how far that kind of record can be trusted, and on what terms.

Who Does What

The two organizations have split the work cleanly. Radian Forge manufactures the test components, collects the manufacturing and sensor data, and conducts applicable testing. ABS provides independent technical review, evaluates the test results and assesses the proposed methodology and its potential application.

That split matters because of what a classification society is. Its value to shipowners, insurers and regulators comes from being the independent party that says a design, material or process meets a defined standard. If the company producing the parts also judged them, the exercise would prove little. Here the manufacturer generates the evidence and the class society decides whether that evidence, and the method behind it, holds up. ABS says it supports the adoption of additive manufacturing through technical requirements and guidance, including its Requirements for Additive Manufacturing, so the project plausibly fits within that existing body of work rather than starting a parallel one.

Michael Kei, ABS vice president of technology, said in the announcement that digital qualification has an important role to play in the adoption of additive manufacturing across the marine and offshore sectors. Jeff Yeager, Radian Forge's chief executive, said traditional manufacturing timelines can create significant challenges when critical components are needed. That is the commercial point of the project: if qualification can lean on data gathered during production, the time between building a part and getting it approved could shrink.

Why WAAM, and Why Ships

WAAM builds metal parts by depositing welded wire layer by layer. ABS frames the interest in additive manufacturing around component availability and long production lead times in the maritime industry, and confidence in the consistency and performance of manufactured parts is, in its words, critical.

Advertisement

Radian Forge's positioning matches that. The company describes itself as combining industrial robotics, WAAM and digital process monitoring to provide an alternative manufacturing pathway for critical metal components for America's maritime and defense industrial base. That last element, digital process monitoring, is what this JDP depends on. A shop that doesn't instrument its builds has no sensor record to qualify against. A shop that does may be able to turn that record into something a class society accepts.

The company's media page shows a young firm moving quickly. Besides the November 12 ABS collaboration, it lists a September 15, 2026 PR Newswire release on a $2.5 million seed round led by Veteran Ventures Capital, and a February 25, 2026 Virginia Economic Development Partnership announcement of a $10.5 million investment in Portsmouth. The page also carries a video titled as building in Portsmouth, next to the fleet the company serves.

Advertisement

The Hard Part: Trusting the Record

The difficulty in data-driven qualification is not collecting data. Modern deposition systems can log a great deal. The difficulty is showing that a given pattern in that data reliably matches a given quality in the finished metal, and that a deviation in the data reliably flags a defect. Until that link is shown, a sensor trace is just a trace.

That is why this project pairs sensor data with inspection records and material test results instead of treating the sensors as enough on their own. To validate a method like this, you build parts, record everything, test those parts conventionally, and then check whether the digital record would have predicted what the tests found. ABS's role in assessing the methodology is the safeguard: a class society has to be convinced the approach holds up before it lets the approach carry any weight.

The announcements don't give a timeline for results, the specific alloys or part types under study, or what would count as success. Those are the details to watch for when the project reports.

What It Means for Makers

Most readers are not printing ship components, but the question this project asks reaches well beyond shipyards. Anyone who has tried to move a printed part from prototype to production has run into the same wall: the part looks right, but proving it is right costs time and money, often by testing or destroying samples.

The direction here is toward treating the build log as part of the product. For hobbyists, that already shows up in small ways, such as keeping print logs, watching first layers with a camera, and recording settings alongside results. At industrial scale, the same habit becomes formal evidence. If a body like ABS concludes that well-instrumented WAAM builds can support qualification, that would be a signal that process monitoring is moving from a nice extra to an expectation for printed parts that have to carry real loads.

The practical takeaway is simple. Record what your machine did, not just what came out. In regulated industries that record may increasingly matter to whether a printed part ships, and the habits that make it trustworthy are the same at any scale.

Sources