How does metal 3D printing work, and where is it used?

Metal 3D printing, more formally metal additive manufacturing, builds a metal part one layer at a time. Manufacturers say it can also combine several pieces that would otherwise be joined into one.

The clearest example in the sources is aerospace: a jet engine fuel nozzle tip that GE Aviation was printing in 2018 at its plant in Auburn, Alabama, described in the company's 2018 announcement. GE's figures for that part are set out under "What is only reported" below. The rest of this article explains the method, separates what the evidence supports from what is only claimed, and sets out what is still unresolved.

How it works

The metal process the sources checked describe most closely is powder bed fusion, in which a laser scans over a thin layer of metal powder to fuse it. GE's announcement describes its parts as grown from a CAD file using layers of fine metal powder and an electron beam or laser.

A DSIAC technical inquiry posted in 2024 illustrates this with a photo caption describing a high-power laser spot scanning back and forth over a layer of cobalt-chrome powder on NIST's powder bed fusion machine. The sources checked do not describe the full build cycle, such as how the platform moves or how layers bond, so those steps are not set out here.

Directed energy deposition, the other main family of metal additive processes, is named in NISTIR 8538 (see below), but the sources checked for this article do not explain how it works, so it is not described here.

What has been demonstrated

"Demonstrated" here means shown in research or government work. It does not mean shown to be routine across industry.

• Process testbeds. NIST has built a laboratory testbed for studying how a laser interacts with metal powder in laser powder bed fusion. Its 2025 report, NIST AMS 100-66, says that controlled conditions and advanced metrology cannot be readily implemented in commercial machines, and that computational models and in-situ monitoring need experimental validation under those conditions. The testbed was built to study the process and calibrate machine performance.
• Monitoring and inspection research. NISTIR 8538, published in September 2024, reviews the current state of the art in methods, instruments and standards for monitoring metal additive manufacturing during fabrication and for non-destructive evaluation carried out during fabrication. It identifies gaps in knowledge, technology and standards. Its summary links the work to a US Nuclear Regulatory Commission effort to certify additively manufactured components for nuclear power plants.

The sources checked do not document established medical or tooling uses, so this article does not claim them.

What is only reported

Some of the strongest claims come from the companies that use the process. The figures below are company statements and have not been independently checked.

• In October 2018, GE Aviation's Auburn plant produced its "30,000th additively-manufactured fuel nozzle tip" for the LEAP engine, according to GE's announcement.
• GE says printing reduced the nozzle tip from about 20 pieces, previously welded and brazed together, to a single piece.
• GE says the printed tip's weight was cut by about 25 percent compared with the earlier assembly.

These are figures from a single 2018 announcement. The page gives no production volumes after that date, and this article does not claim any later figures. The same announcement says parts are made in a fraction of the time of traditional methods, but gives no figures. Claims that printing is faster or cheaper than conventional methods, or that it is ready for volume production of many parts, rest on company statements that the sources checked do not verify.

What is speculative

Several ideas are often discussed but are not shown by the sources. The most common is that metal printing will replace casting or machining for mass production. The sources reviewed give no verified cost or speed comparison with conventional methods, so the economics of any given part cannot be judged from them. Other ideas, such as printed heat exchangers with internal cooling channels, or spare parts printed on site for remote locations, are plausible directions but are not evidenced here.

Limits and open problems

Defects and printability. A 2023 editorial in Materials by Hong Wu and colleagues says that poor printability and unacceptable metallurgical defects are still the primary concerns that limit the adoption of most alloys in the laser powder bed fusion process (Wu et al., 2023). The sources checked do not say which defects these are or measure how far they reduce part strength, so that effect is not stated here.

Monitoring and inspection. Checking that every printed part is sound is still an active research area. NISTIR 8538 is devoted to this problem and reports gaps in knowledge, technology and standards.

Qualification and standards. The DSIAC inquiry, posted in March 2024 (completed in August 2023), reports that outside NASA standards there is no DoD-approved qualification and certification process for additive manufacturing. It also says that additive manufacturing is not as trusted across the US Services and DoD as traditional manufacturing processes. Qualification is the process of showing that a part, process or machine is reliable enough for a given use.

Cost and speed. The sources give no verified comparison of cost per part or build time against conventional manufacturing.

Materials. Because the Wu editorial says poor printability still limits most alloys in laser powder bed fusion, the range of metals that can be printed reliably by that process remains limited.

What to watch next

• Whether NIST's testbed and monitoring work leads to measurement methods and standards that qualification bodies accept.
• Whether the nuclear certification work referenced in NISTIR 8538 produces a published route for printed components.
• Whether defence qualification gaps highlighted by DSIAC are filled by agreed, published processes.
• Whether manufacturers publish production volumes, costs and inspection results that can be checked independently.