AM has not been immune to the emerging tech hype cycle. High hopes that 3D printers would become household appliances in the early 2010s fizzled on account of AM’s limitations at the time, leaving a scar for some investors.

But the technology has made considerable progress since then, proving competitive for low-volume production and mass customisation of industrial-grade parts. Manufacturers are using it to develop custom tooling and assembly aids at factories, while the medical industry is turning to AM to print implants, prosthetics and other devices tailored to patient needs.

Still, AM faces major hurdles to broader adoption for mass production. Manufacturers can achieve greater economies of scale using traditional manufacturing methods, and AM poses unique reliability challenges because it often entails creating new parts and new materials at the same time. As a result, where mass production is concerned, AM has been largely relegated to printing complex, high-value parts in select industries, such as aerospace.

In the most likely scenario going forward, AM will serve as a complement to traditional manufacturing, rather than replace it.

What may be overlooked is the role that artificial intelligence can play in increasing AM adoption as it matures. Much of AM’s progress today is based on trial and error, but machine learning could eventually give 3D printers “eyes” (machine vision) and “brains” (closed-looped feedback) to help advance the AM process and improve cost, speed and reliability faster than anticipated.