The Star Trek replicator stands as the ultimate dream of on-demand manufacturing. The bad news is that despite advances in 3D printing we are many breakthroughs (including nearly abundant energy) away from realizing that dream. The good news is that we are finally finding ourselves on the cusp of a paradigm shift in manufacturing away from mass production to fast, low cost, small series (including single part) capabilities.
The shift is being led by a new crop of startups, including Plethora (metal parts), PCB:NG (circuit boards), Unmade (knitting) and Shapeways (3D printing). The common theme among these is that they are focused on automating pre- and post-production processes. For instance, Plethora has created software for translating a CAD drawing directly into the right sequence of instructions for a CNC Mill. PCB:NG has figured out how to feed a pick and place machine for runs of just a few boards. And Unmade can drive a knitting machine right from the design of a scarf or sweater.
In 3D printing, our portfolio company Shapeways has invested massively in making pre- and post production more efficient. For instance, in pre-production, there is a task called tray planning, which involves figuring out which parts should be printed together in one print run and how those parts should be oriented and combined to achieve high density. In post-production, there are tasks such as breaking parts out of the block of material, cleaning, dying and packing for shipment.
Why is this happening now? On the demand side, the long tail has been growing driven by consumers who want customized products, hobbyists and makers who want custom parts and innovative hardware startups. This demand growth is absolutely critical as without it there is no payoff to automating pre- and post-production. On large production runs the pre- and post-production cost gets amortized over a large number of identical parts and so is irrelevant to the cost per part.
On the supply side, the core manufacturing machines have become completely computer controllable. Switching seamlessly from one part to a completely different one next is straightforward on a CNC mill and even on a modern knitting machine. But what has remained hard is figuring out how to go from the design to the code driving the machine, getting the right supplies into the machine and removing, finishing and shipping parts.
The OEMs of mills, pick and place machines, knitting machines, and even 3D printers have done next to nothing to address pre- and post-production. That’s because until recently these machines were used by companies that were largely mass manufacturers. The automation of pre- and post-production is now getting cheaper through improvements in software and also advances in robotics. For instance, optical recognition of parts is going from costly and error prone to being achievable with low cost cameras and off the shelf hardware.
Plethora, PCB:NG, Unmade, Shapeways and others represent the leading edge of what is possible. To date they account for a tiny fraction of all manufacturing. But I am excited that they are showing the way past mass manufacturing to a future where stuff is made only when someone actually orders it, is customized to their needs and is produced in geographic proximity (reducing shipping costs).
