From Geometry to Process Control: What Limits Industrial Metal AM — Interview with Amin S. Azar of 3D-Components

3D-Components AS is a Norwegian technology company developing solutions for industrial metal additive manufacturing. Its focus is on robot-based, data-driven approaches aimed at improving process stability and reproducibility, particularly in industrial applications with high demands on scalability and traceability.

Rather than treating 3D printing as a static sequence of fixed parameters, the company approaches additive manufacturing as a dynamic process. This approach is implemented in RobTrack, 3D-Components’ primary AI-based automation solution for robotic welding and additive manufacturing. RobTrack combines machine data, sensor inputs, and models to monitor and adjust production conditions during deposition, with the goal of increasing repeatability and predictability while reducing material waste.

Process qualification remains a central hurdle for the industrial use of metal additive manufacturing, especially in regulated sectors. 3D-Components addresses this challenge by focusing on validation of the manufacturing process itself, linking real-time production data with analytical models to support traceability while reducing the need for extensive post-process inspection.

Alongside software development, the company investigates material behavior under the thermal conditions typical of metal additive manufacturing. By correlating process data with material properties, 3D-Components aims to enable more predictable and consistent use of additive manufacturing in industrial production environments.

Interview with Amin S. Azar

In the interview, founder and CEO Amin S. Azar discusses how industrial metal additive manufacturing is evolving from experimental use toward data-driven, automated production environments. He shares observations on current limitations, realistic expectations, and how advances in process control, robotics, and AI could shape qualification, scalability, and adoption across the broader AM industry.

How do you see the role of AI and closed-loop control evolving in industrial metal additive manufacturing, and what technical barriers still prevent fully autonomous production today?

AI models are transitioning from offline decision-support systems focused solely on recommendations to proactive and reliable systems for monitoring and correction. Whereas most current AI-enabled workflows identify anomalies after they occur, future advancements point toward predictive control, where AI anticipates potential deviations and dynamically adjusts parameters to prevent issues before they arise. This progression is especially significant for minimizing scrap rates in high-value industries such as energy and aerospace, as well as optimizing operational efficiency in manufacturing industries.

The goal of fully autonomous production faces challenges related to data latency, the availability of high-quality datasets, and rigorous verification protocols. Processing vast quantities of real-time sensor data rapidly enough to enable parameter adjustments necessitates advanced edge computing solutions that are still under development. Additionally, in highly regulated sectors, relying on “black box” AI systems for autonomous decision-making introduces complexities in certification and auditing. To address these concerns, there is a pressing need for explainable AI models and standardized frameworks that validate AI-controlled processes against the stringent reliability standards of traditionally engineered operations. Our team at 3D-Components is strategically positioning itself within the market, with an exclusive commitment to developing AI tools focused on robotic directed energy deposition (DED).

Based on your experience in the field, which expectations around industrial 3D printing have not materialized in real manufacturing environments, and what should engineers and decision-makers learn from that?

The notion that increased complexity comes at no cost, along with the vision of effortless “push-button” manufacturing, has not realized its anticipated potential. Although additive manufacturing (AM) is highly effective for producing intricate geometries, the industry has often underestimated the expenses and challenges associated with pre- and post-processing steps, which can account for 40–60% of the total cost of a part and may offset the advantages gained during the printing phase in various business scenarios.

Furthermore, approaches such as “digital qualification,” “first-time-right,” the currently subjective “design for AM,” and “through-build properties control” remain challenging to implement due to the complex and concurrent multiphysics nature inherent in additive manufacturing processes. This presents an opportunity for artificial intelligence, which excels at managing and optimizing operations within such high-dimensional processing environments. We have recognized these needs and consolidated various AI-enabled solutions into our RobTrack system, which is designed to provide users with a comprehensive tool that allows them to fully leverage the capabilities of their DED systems.

What aspects of materials behavior and process stability in metal AM are currently the least understood, and where do you believe the next major breakthroughs will come from?

Within the field of metal additive manufacturing (AM), microstructural evolution during nonequilibrium cooling—particularly as it pertains to the interaction of the heat source, melt pool, and repetitive thermal cycling—remains one of the least understood phenomena. The industry continues to encounter challenges when attempting to correlate material properties with processing conditions. Variations in thermal history, which are strongly influenced by geometry, can lead to inconsistencies in properties across a single component. Moreover, each alloy, along with its unique chemical composition, presents specific complexities in establishing optimal processing parameters.

Future advancements are anticipated through artificial intelligence, enabling the training of extensive models that incorporate a broad spectrum of physical and mechanical properties to rapidly generate reliable and material-specific processing conditions. Although this is a resource-intensive endeavor, its potential benefits are substantial. It is essential for industry to recognize these opportunities and allocate investments accordingly. Delays in such investment may hinder technological maturation and impede growth within the additive manufacturing sector.

Looking ahead five years, how do you expect robotics, data-driven manufacturing, and advanced simulation to change how additive manufacturing systems are qualified, scaled, and adopted across industry?

Over the next five years, emphasis in advanced manufacturing will transition from “qualifying the part” to “qualifying the process,” also referred to as “part-agnostic qualification.” Leveraging sophisticated AI models and digital twins will enable virtual validation of builds prior to any processing, significantly reducing reliance on physical trial and error. Rather than conducting destructive testing for each build, data-driven “digital passports,” grounded in a qualified process, will serve as verifiable evidence of quality.

Concurrently, advancements in robotics will address the existing automation gap. Engineers will be able to swiftly configure process parameters and define optimal tool paths tailored to specific materials and machines. A clearer delineation between machine operations and operator responsibilities will emerge, elevating human–machine collaboration to new heights. This evolution is critical for scalability, transforming additive manufacturing from a niche laboratory procedure into a continuous and reliable factory-floor production system capable of sustained 24/7 operation.

Our flagship product, “RobTrack,” aligns closely with this outlook, and our system is designed to help users realize these promises in the future of robotic welding and directed energy deposition technologies.

Further information on 3D-Components can be found on the company’s website.