BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Building the New Manufacturing Stack

Following
This article is more than 4 years old.

What does the new "manufacturing stack" look like?

We’re all familiar with the Internet’s TCP/IP stack. The bottom layer handles the details of the physical transmission of raw bits and bytes, and applications ride on the top layer of the stack. Meanwhile, the intermediate layers glue it all together—breaking down data into packets, routing it reliably and quickly to its destination, and reassembling the packets into the original data.

This model has turned out to be highly flexible, offering new opportunities for innovation and productivity that were not present in the old point-to-point model. The TCP/IP stack allows each layer to do its job without knowing the details of adjacent layers.  

Manufacturing is going through a similar transition, from a linear “supply-chain” world to a much more adaptable “Internet of Goods.”  And just like the Internet is built on the TCP/IP stack, the Internet of Goods is going to be built on a new manufacturing stack that goes beyond the factory floor.

The figure below lays out our model of the new manufacturing stack. The bottom layer of the manufacturing stack, layer 1, is the physical movement of raw materials, parts and other intermediate goods, and finished products. Layer 2 focuses on factory level activities, such as the production and transformation of goods. Layer 3 includes the assessment of manufacturability, estimates of “best case” cost, and routing to the appropriate production node. And Layer 4 is about designing the product.


Much of the new manufacturing stack is already in place. Layer 1 encompasses logistics, the art and science of getting goods to where they are going, quickly and cheaply. Obviously this is not new, though Amazon has made tremendous strides in increasing the productivity of warehouses and the fulfillment process. Layer 2 focuses on factory level productivity, including the growing use of robots and IT systems that give a comprehensive view of the entire factory operation. Layer 4, the separation of design from production, has been practiced for years as ‘fabless manufacturing” or simply outsourcing of production.

What’s new is Layer 3. In the past, it’s been difficult to separate out assessments of cost and manufacturability from the actual production process. Manufacturers would quote potential buyers a price, based on their own understanding of their costs and market forces. Or large companies would work with suppliers to hash out a design and a price.

But now manufacturing platforms such as Maryland-based Xometry, Amsterdam-based 3D Hubs and San Francisco-based Fictiv are using machine learning to accumulate a growing mountain of data about “best case” costs and manufacturability for parts.

This data did not exist before, and it opens up all sorts of new business models. For one, the manufacturing platforms have the capability to take advantage of underutilized manufacturing capacity, just like AirBnB takes advantage of unused living space. ”You don't need to build more factories,” says Fictiv CEO Dave Evans, ”you just need to better utilize the existing ones.”

This model, which is sometimes called distributed or on-demand manufacturing, is going great guns. Fictiv has produced over 10M custom mechanical parts for both prototype and production across all of the major industry verticals. At 3D Hubs, revenue is growing 140% year-on-year, and the company has just added teams in Chicago, Shenzhen, London, Paris and Berlin. Xometry has continued its annual revenue growth of over 100%, says CEO Randy Altschuler, and is looking to launch in Europe later this year.All three companies have recently attracted big venture capital infusions. In 2019 alone 3D Hubs has raised $18 million, Fictiv $33 million, and Xometry $55 million.

The manufacturing platforms are attractive investments because they are collecting data that attacks the fundamental inefficiency in manufacturing: A lack of shared knowledge about the lowest-cost ways of making things. That may seem odd, but economic research consistently shows that the most productive plants make almost twice as much output with the same measured inputs as the least productive plants, with even larger productivity differences in China and India. University of Chicago economist Chad Syverson writes: “Put simply, some producers seem to have figured out their business (or at least are on their way), while others are woefully lacking.”

Consistent with these results, 3D Hubs found that supplier bids for a project sometimes differed by a factor of 3. “We could not clarify why there is such a price difference,” said Bram de Zwart, CEO of 3D Hubs.

Instead, says de Zwart, “we basically teach our manufacturing partners what is the right price.“ In effect, the manufacturing platforms are building up databases of what the “production frontier” in manufacturing looks like, which can then be transmitted to lagging producers.  "Our software must analyze part geometry and open capacity to find a market-efficient price that benefits both our customers and our manufacturing partners," says Altschuler of Xometry. Adds Evans of Fictiv: "Our core IP lies in our computational geometry analysis. We analyze CAD data to predictively calculate true cost, machine time, and manufacturing feedback."

Another benefit of the manufacturing platforms: Designers can tap into the databases of these companies as they refine the product design. "We can tell them if those iterations make sense in a pricing and manufacturability sense," says de Zwart.

Finally, and perhaps most important, the trade war between the United States and China is making companies more open to new ways of doing business. "With all of the tension in global trade, we're seeing a real need for agility in supply chains," says Evans of Fictiv. "Companies are coming to us to better understand how they can ramp production during a trade war." In the event of an all-out trade war between the United States and China, adds de Zwart, "we would direct those orders to more local suppliers."

The new manufacturing stack has the potential to unleash creativity in unexpected places. In Murcia, Spain, a new online manufacturing platform called Faberin is bringing together designers with local master craftsmen to offer furniture manufactured on demand for the end consumer. This could invigorate the markets for local designers and artisans.

India is home to new manufacturing platforms as well, including Pune-based Chizel, and Bangalore-based Zetwerk. These innovative enterprises could boost India's small and medium size manufacturing sector.

The new manufacturing stack could unlock as much creativity and growth in the industrial sector as the TCP/IP stack did for the Internet—not just in the advanced countries but around the world.

Follow me on Twitter or LinkedIn