At Desktop Metal, on a wild unicorn ride with Ric Fulop (and Ely Sachs)

On a recent trip to Boston, to co-host and present at the AM Strategies conference on Medical and Dental AM, I was invited to visit Desktop Metal. The fact that the company just closed the latest round of investments, for $160 million (bringing the total to $438 million) made it all the more appealing. Desktop Metal is one of the companies that could truly change the face of additive manufacturing, taking over industrial mass production. Seeing up close how (and if they really) intend to do this is an opportunity that I simply could not miss.

I rented a car, beat the Boston morning traffic, and headed out to Burlington, where the company is based. Lynda McKinney, the company’s Head of Global Communications told me that the company’s founder, Ric Fulop, would be there to show me around. I knew Mr. Fulop was one of the key drivers for Desktop Metal’s vision but I honestly had no idea at the time how extensive Mr. Fulop’s influence has been on several Boston companies that are shaping the world of digital and additive manufacturing.

Before founding Desktop Metal, Mr. Fulop was a General Partner at North Bridge. He led the Series A rounds in, among others, Onshape and MarkForged. In addition, North Bridge was the original Series A investor in several of the companies that have changed the 3D and CAD/CAM landscape including SolidWorks (acquired by Dassault Systemes), ProtoLabs (now listed on the NASDAQ), SpaceClaim (acquired by Ansys) and Revit (acquired by Autodesk). In 2001, he founded A123 Systems, to commercialize technology developed at the Material Sciences and Engineering Department of MIT. At his departure, the company had a public market capitalization of $1.5 billion, 1,800 employees, with production facilities around the world and had shipped hundreds of millions of dollars in energy storage systems (i.e. the batteries that are shaping the future of mobility).

Welcome to Desktop Metal

The Desktop Metal facility is located in Burlington, about 15 miles outside Boston. In spite of the freezing temperatures caused by the Polar Vortex, the Desktop Metal facility is warm, a welcoming and ultra-modern place to work in. It is mainly a large open space, with some separate sections for meetings and R&D. It somewhat reminds me of the huge open space I saw at HP’s R&D facility in the Barcelona office. As I discover later in my visit, the similarity is not entirely casual: this facility is mainly dedicated to engineering research. Over 300 engineers are working every day on improving Desktop Metal’s binder jetting technology. In doing this they can also benefit from key contributions by none other than the inventor of binder jetting, MIT professor Ely Sachs.

I catch up with Mr. Fulup and, as we start walking around, the first question that comes to mind is: why did investors give $160 million to Desktop Metal and not to another, possibly more established 3D printing company? I ask Mr. Fulop but he doesn’t immediately answer, saying that he “doesn’t understand the question” and offering a rather cryptical “investors give money to make money”. The answer would become self-evident by the end of my visit.

To him, it is probably obvious but I still struggled to understand why investors are so confident that Desktop Metal’s approach is going to be the winning one (or “a” winning one). I also realize early on that Mr. Fulop is not a man of many words (although this could also be due to his laryngitis). Like many successful businessmen, he seems more interested in making progress and expanding his company rather than talking about it.  While he had a key role in getting Onshape, Protolabs and MarkForged off the ground, Desktop Metal is clearly the company where he sees the greatest possibilities. What I want to understand is why.

The half a billion dollar plan

We go through the various sections of the open space as Mr. Fulop points out how streamlined operations are. First, we pass by the area where the debinding systems – one of the hardware systems that complement the Studio System’s workflow, together with the sintering furnace –  are developed and tested. He points out that this is the World’s only debinding system that can work in an office. There are no vents. All you need to do is plug it and you can run it from your iPhone. This is the appetizer to Desktop Metal’s vision for the future of manufacturing. The system does all the work. The Production System – I find out shortly thereafter – is even more streamlined as it requires no debinding at all. “We go straight to brown parts and we put them directly in the sintering furnace,” Mr. Fulop tells me. This, I imagine, will eventually be carried out automatically, using robotic carts to transport the build plates to and from the systems. The telepresence robot I see in one corner further strengthens this idea.

We proceed through the different stations, where several metal parts – green and finished – are laid out. “We can make very big parts as you can see,” Mr. Fulop says. “However we don’t do any production services at all. We only make machines and have no intention of competing with our clients and suppliers.” As we pass by the office of the CRO I see Tuan Tranpham in a meeting. It reminds of me of the time when he announced he was leaving Arcam, just before GE’s acquisition. Then it could have seemed like a bold move but he clearly was convinced by Desktop Metal’s vision from the onset. Time seems to be proving him right, although GE is also continuing to make its own big moves in this market (with an eye on binder jetting as well). At the very least, though, Desktop Metal is going to compete, putting the company in a position not so different from Tesla vs. the large automakers. Considering that Mr. Fulop built most of his capital with an electric car battery company, the comparison seems quite befitting.

At the very least, Desktop Metal is going to compete, putting the company in a position not so different from Tesla vs. large automakers. Considering that Mr. Fulop built most of his capital with an electric car battery company, the comparison seems quite befitting.

Laying down the foundations of a new industry

Like Tesla’s Elon Musk, Ric Fulop is probably the most “serial entrepreneur” in 3D printing. Like Musk he made a lot of money by selling his equity in a huge company he contributed to build and set out to change the world doing what most considered impossible at the time: accelerating metal 3D printing to the point that it could compete with all other production processes, both formative and subtractive. In doing this he did not just target metal AM but also diversified targeting other segments.

“We put the money to start Protolabs and owned a third of it. It was a small prototyping service and we gave them the money to scale up and make it real,” he reveals. “When I left A123, the company had 70% market share in start-stop, energy storage systems for vehicles, supplying most of the major automakers. In 2010 I wanted to do something new so I sold my equity in the company that had reached $1.5 billion in market cap. From 2010 to 2015 I invested about 130 million dollars in eight companies. As I got involved in Protolabs, I wrote a thesis on investing opportunities in the 3D space. I theorized that someone should do composites,  someone should do elastomers, someone should do high-speed metal printing and CAD in the cloud. So I started looking for opportunities.”

I was an early investor in Onshape and gave Greg Mark the money to start Markforged, as I continued to make investments in the 3D space.

Among all these Mr. Fulop chose to focus primarily on Desktop Metal, a company which he built together with several high-profile co-founders. They include MIT Professors Ely Sachs (Ph.D.), the inventor of binder jetting technology, and John Hart, who leads MIT’s industrial additive manufacturing activities. They also include Mr. Jonah Myerberg, as Chief Technolgy Officer, whom we’ve interviewed some time ago.

“At Protolabs, we were at the time the second largest customer for Concept Laser,” Mr. Fulop continues. “We owned 12 Concept Laser machines. If you think about it, that was not a very large number to be the second largest customer of someone.” Through that experience, Fulop realized that metal PBF was a very challenging process, time-consuming and labor-intensive. “Because you are combining shaping of the part with thermal processing it reduces the speed, which is made even slower by the need to remove supports, making it very hard to scale.”

Just as we discuss the challenges of PBF technology we run into Prof. Ely Sachs. He is in the middle of something important, moving fast across the “factory” floor.  Mr. Fulop says that he is the “busiest man he knows”, which kind of confirms the stereotypical image that we have of geniuses who will go down in history for their inventions. Ely Sachs is likely one of those men but he took the time to stop and say hello and even pose for a photo (the only photo that I was allowed to take, with Mr. Fulop making sure that I use the portrait function to close in on them and not on other details). Mr. Fulop’s introduction went something like this:

“Davide, here, is an industry analyst and runs a series of publications on 3D printing. Ely invented binder jetting… “

History in the making

As we make our way through the company we run into important new Desktop Metal clients and investors coming to see the systems at work. I am kindly asked to stand aside to avoid listening in on non-disclosable conversations. I do get the feeling that I am witnessing history in the making, something that – I admit – is fairly common when I visit important 3D printing companies.

Mr. Fulop agrees that the AM industry is really evolving at a fast pace and gives me some numbers to show how Desktop Metal will address and channel this evolution. “In the world of laser metal powder bed fusion, you have one million dollar capital equipment cost for one metric ton a year of production,” he explains. “By separating the thermal processing from the shaping of the part with our single jet passing technology we are able to do 150 metric tons of production for one million dollars of cap-ex. That makes it over 100 times faster. In addition, the raw materials for our process are up to 80% cheaper, mainly because our powder is not affected by oxygen content while PBF powder has to be completely oxygen-free.  This makes the powder making process also much more cost-effective. When you combine 100 times faster with 80% cheaper materials you start to become cost-effective against mass production technologies like die-casting on very large batches.”

We are able to do 150 metric tons of production for one million dollars of cap-ex. That makes it over 100 times faster [than PBF]. In addition, the raw materials for our process are up to 80% cheaper.

By this time we have confirmed that Desktop Metal’s version of the metal binder jetting process is significantly faster and more cost-effective than metal PBF. But – to this date – metal binder jetting has been used mainly for prototyping and some very small scale productions, while PBF has already been extensively tested and validated by almost every major standard and regulatory committee, as well as aerospace and medical companies. How long will it take before metal binder jetting is truly ready for adoption? Mr. Fulop thinks not long at all. “The second time around is always faster. In addition, what really matters is that we meet material properties. Committees don’t generally certify specific parts. They want to make sure that the material properties in a specific process meet requirements. Our process does meet these requirements. In PBF the microstructure of materials was something completely new. With our technology, we have the same microstructure of more traditional sintering and casting processes. That is already very well understood.”

Scalable, factory-less manufacturing

The bottom line is that the Studio System is now shipping in large volumes. Mr. Fulop said Desktop Metal has shipped more metal printers in this quarter than all other metal 3D printing companies combined (a phrase that we have also heard from Formlabs in the past, referring to SLA).  “Our main challenge with the Studio System was to finalize the engineering so we could ship in large volumes. On the Production System’s side, we are starting to ship next month the first systems to non-beta clients.” Mr. Fulop revealed that the Production System currently has a  one-year waiting list. “This will decrease significantly by next year as we reach full production capacity,” he adds.

In Davos, a lot of the discussion revolved around Industry 4.0 and additive manufacturing is an integral part of the next industrial revolution, especially as it moves toward full-scale production

Like Airbnb, a hotel company without hotels, or Uber, a car company without cars, Desktop Metal is a hardware manufacturing company without factories. Its 300 engineers work on solutions to scale up production so that large external manufacturers such as Jabil (which also makes HP’s systems) can make and ship the systems faster and more effectively. Several other companies in the 3D printing hardware manufacturing segment follow this model. The main difference is that most of them don’t have 300 engineers constantly working on advancing processes.

This approach is perfectly in line with industry 4.0 dynamics which Mr. Fulop recently discussed at the World Economic Forum in Davos. “The focus on AM has been increasing. In Davos – he says – a lot of the discussion revolved around Industry 4.0 and additive manufacturing is an integral part of the next industrial revolution, especially as it moves toward full-scale production.”

Which are the segments that will adopt Desktop Metal’s technology for production? Will it be aerospace and medical companies that are already familiar with AM in general, or will it be new industries that have higher throughput requirements such as automotive and beyond? Mr. Fulop does not think of it in these terms. “We provide a platform. What companies do with it is not something we overly concern ourselves with. It’s like a computer or a smartphone: we make 3D printers and 3D printing is a horizontal technology. Adopters will decide what the best applications are.”

In terms of productivity, you have to think of them [each Production System] as 100 EOS 400-4 lined up.

What’s next

Desktop Metal is moving fast. It began shipping the Studio System in volumes, it will begin shipping the first Production Systems next quarter and Mr. Fulop anticipates that they will be introducing new systems as well. Evolution could go in several different directions as single pass binder jetting has no inherent limitations in terms of build size and materials (Desktop Metal just introduced the new 316L stainless steel) “Yes, everything,” says Fulop. “One of the most exciting aspects of this process is that we can print existing materials that are not compatible with PBF, such as  D2 Tool steel or 4140 aluminum. We could do refractories, silicon carbide and more.”

To further prove this point Mr. Fulop takes me into a section where two Production Systems are running. I have seen them at shows, however seeing them inside an enclosed space makes them look a lot larger. “You have to think of them as 100 EOS 400-4 in terms of productivity,” Mr. Fulop says to put it into perspective. “How big would the room need to be? In this room, we have two of them, which means 300 metric tonnes a year in parts.”

With that kind of productivity, could these systems one day replace all the CNC’s and die-casting systems in the world? I once spoke with Jabil about the fact that they had 15 HP MJF systems and roughly 15,000 CNC’s around the world. Mr. Fulop tells me that one of Jabil’s competitor has some 300,000 industrial CNC systems installed. I conclude that there is no reason why not, which by extension would indicate an unprecedented potential for growth in AM.

And that, incidentally, also answers my initial question on why investors would bet half a billion on Desktop Metal.