The Next Era Of Additive Manufacturing Is Trusted Production

Dr. Ali Tamijani is the Founder and CEO of Novineer, an AI engineering software company for reverse engineering, simulation, and design.

getty​Additive manufacturing has proved it can make parts. It hasn’t proved, at scale, that it can produce parts that meet requirements, pass certification and still make economic sense. That gap between a flexible fabrication method and a trusted production system is what the industry must now close.​

Global additive manufacturing revenue reached $24.2 billion in 2025, according to Wohlers Report 2026, but annual growth has slowed to 10.9%, down from the 20%-plus highs. Printing services, the companies actually making parts, now account for 48% of industry revenue and grew 15.5% in 2025, while new system sales grew just 3.6%.

Value is concentrating in production rather than hardware. This means the industry’s future no longer depends on selling more machines but on whether the parts coming off those machines can be trusted.

For safety-critical applications, trust is necessarily expensive to establish. By NIST’s estimate, fully qualifying a new material often demands many thousands of individual tests, millions of dollars and between five and 15 years. Faced with that burden, many manufacturers keep additive in the prototyping lab, where mistakes are cheap and nobody asks too many questions about a part’s pedigree.​

Disconnection causes the qualification burden. Today’s printers are already more capable than the workflows that surround them, but trusted production requires a workflow in which every stage informs the next: design to process definition, build, inspection, performance and cost. In most organizations, those stages live in disconnected tools, and every disconnect is a place where trust leaks out.​

If you simulate a part made by material extrusion and predict digitally how it will perform in service without knowing its actual toolpath, you’re predicting the behavior of a part that may never be built. When the digital models can’t be trusted, the only alternative is to break real parts by brute-force physical testing until the statistics say stop.​

The industry has been promising a “digital thread” for additive (and manufacturing, more generally) for more than a decade, so why does it still not truly exist? The honest answer is that the connections we built were file-based, not physics-based.​

A design or simulation software passes a design file to a print-preparation tool, and the print-preparation tool hands a toolpath to the machine. The tools exchange files, but understanding the physics behind those files is a separate problem that neither tool solves. A workflow is truly connected only when it can answer the question that matters on a production floor: If I change X, what happens to Y?​

To see why, it helps to understand what these files actually contain. A CAD model describes the ideal shape, the dimensions and the tolerances. It says nothing about how to make the part. Before printing, the geometry is sliced into thin layers, and each layer is converted into a toolpath: the machine’s actual choreography, specifying where nozzle travels, in what order at what speed and temperature. The blueprint isn’t the recipe, and each translation between them loses information.

By the time the machine is running, it’s executing millions of movements that no longer carry any record of the requirement they serve. In additive manufacturing, even the material itself doesn’t truly exist until the part is sliced.​

In conventional manufacturing, this disconnection was tolerable. A machined bracket begins as a certified billet, its material properties managed separately to geometry. With additive manufacturing, the material is created at the same moment as the geometry, and the toolpath is the instrument that creates both.

Two parts with identical geometry and identical feedstock, printed on the same machine and loaded the same way, can differ in deflection by a factor of six because of nothing but the toolpaths used to fill them. In additive manufacturing, the process is the material, and any workflow that separates the two will be structurally wrong.​

A printed part has an anatomy: its geometry. It has a physiology: its material. It also has something most discussions overlook: a nervous system. The toolpath choreographs how the part comes into being, layer by layer, and together with the thermal history, it determines how the part will perform under load.​

Binding it all together is memory: the digital thread that records how everything came together in a specific build so the result can be reproduced, inspected and trusted. Connected workflow generates that memory, allowing a part to be certified and reproduced with confidence. A part without it can only be repeatedly retested.​

Today’s qualification regime copes with uncertainty by attempting to freeze the machine and process parameters. Any change restarts the testing and increases cost. When models validated against physical test data are connected to the actual toolpath and build record, evidence can transfer from one part to the next, and qualification becomes an asset that compounds rather than a cost that repeats.​

1. Map where design, process, simulation, build and inspection data live today, and identify every point where files are transferred rather than underlying physics. Each of those gaps is a future qualification cost.

2. Treat the toolpath as first-class engineering data. If the toolpath determines performance, it belongs under the same version control, review and analysis discipline as geometry and material.

3. Demand memory from your software and your suppliers. Every production part should carry a record of how it was designed, built and verified. Anything less is someone selling you a prototyping machine at production prices.

​Production won’t forgive parts without memory or a nervous system. This leaves a question every manufacturing leader will eventually have to answer: When a customer asks for the record behind a part, will you have one to hand them?​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?