How GE Aviation Avoided a 3D Printing Innovator's Dilemma

GE Aviation and and 3D Printing Supremacy

There was a headline recently that might have gone unnoticed:

“GE Aerospace to Invest Another $1B in U.S. Manufacturing”

From the press release: The investment expands capacity at sites producing and assembling commercial and defense engines. This includes $115M in Cincinnati, Ohio—home to GE Aerospace’s headquarters— to modernize infrastructure, increase test cell capacity, and expand advanced 3D metal printing capabilities.

Buried in there is normal stuff - modernizing infrastructure, increase capacity, etc. But there is also something that always sticks out whenever I am reminded of it: GE is a 3D printing giant.

There was a point in time when 3D printing (at some point rebranded to additive manufacturing) was the current thing in venture capital. The idea was that we would 3D print everything - homes, cars, guns, ladders, any household need, we could just load up the 3D printer with the right materials and, voila, whatever you needed was there.

But that didn’t quite come to fruition. We got some interesting 3D printing outputs. There were a lot of successes and failures along the way - Carbon, Desktop Metal, Markforged, etc. But the consumer wave never really took hold. Most people don’t design parts on their own. And most people don’t need custom parts for everything that they do. Maybe this is a software problem or an education problem or a marketing problem - but regardless, it just never really happened.

But it did happen for industrial companies, particularly in the Aerospace industry. Not to the extent that the VC hype cycle would lead you to believe - completely replacing traditional manufacturing as we know it. But it has certainly had an impact.

SpaceX, for instance, spends a ton of money on 3D printing. The SuperDraco (image below) was famously one of the first fully 3D-printed rocket engines to actually fly. Falcon 9 flew with 3D printed engines as early as 2014 and large portions of the Raptor engine rely heavily on additive manufacturing. Musk has stated that there are parts of the SpaceX manufacturing process that would be impossible without 3D printing - the parts they make are so complex and technically impossible to build without additive features. Even NASA has a 3D printer aboard the ISS.

But that’s SpaceX, maybe the most innovative company in the world, helmed by one of the most technologically-speaking forward-thinking people to ever live. They have built SpaceX from a classic startup framework of taking risks and moving quickly. Fail fast, learn from those failures, and iterate. Don’t be bogged down by existing supply chain relationships because you don’t have any. Use first principals to build better.

Most of us mere mortals don’t have the wherewithal to build at that level of pace. They also don’t have the balance sheet that SpaceX has. Nor does SpaceX have the incumbent forces working against them that most Aerospace companies do. SpaceX has an insane about of capital and investor interest to pile in more cash. So when one of their rockets blow up, it’s not a huge deal so long as it drives the overall story forward through learning and development. They can afford it, both literally and metaphorically.

GE Aviation, on the other hand, does not have that luxury. It has less capital and investor interest, less margin of error (a spaceship blowing up is one thing, but a jet engine blowing up is a whole other can of worms - the regulations are immense), and they have to deal with a classic innovator’s dilemma of an embedded supplier base that relies on them for certain parts.

If you have ever spent any time around GE Aviation, you know that the folks who run that place are the engineers and the supply chain guys. The complex web of manufacturing a jet engine is astounding for the uninitiated. It makes auto manufacturing look like a preschooler playing in the school sandpit. They have hordes of people focused on improving margin 10 bps through supply chain and design know-how.

And imagine you are GE: you have world class suppliers across casting, precision machining, and forging, as well as decades of process knowledge and billions invested in tooling and engineers. Then rolls along 3D printing. It was initially more expensive, slower, lower throughput, less proven and difficult to certify. These are all classic innovator’s dilemma warning signs. The incumbents internal leaders would ignore 3D printing because it would go against their immediate term interests: reducing margin and designing better product. They would be losing money on a product that doesn’t work as well and anybody internally would likely be putting a bonus or reputation on the line for something with little upside.

Classic innovator’s dilemma. So what do you do?

It’s hard to know exactly what GE’s response was internally, whether it was a knock-down, drag-out fight between the forces of good and evil (or innovation vs stagnation), but it’s clear that they answered the bell, eventually, either way. Instead of viewing 3D printing technology as a way to make the same part cheaper (classic incumbent behavior would be to ask how the new technology can make the current product better), they instead asked how the new technology could help them design better parts altogether. Instead of focusing on how to improve margins, which most companies’ mid-level managers are focused on, GE focused on design and just making better products.

But how do you even define “better products”? The aerospace industry has a lot of vectors for how they think about better, but one that is always top of mind is fuel efficiency. If a new manufacturing method improves weight, reliability, fuel burn, etc. then airlines spend less on fuel (one of their biggest expenses and the reason they are so hard to run). As a result, they would be more willing to buy that product (the plane, not the jet engine, but it’s hand-in-hand) and the willingness improves GE’s economics broadly.

GE could have still fallen into this trap as the actual selling of the engine is not really where they make their money, but in servicing of an engine, where most of the margin is. If additive parts reduce the amount of servicing needed (because they are more reliable, etc.), then they might cannibalize a core income stream. But the thing is, aerospace industry is very clear on one thing: fuel efficiency rules. And you can’t service a jet engine that doesn’t get sold in the first place because it isn’t very fuel efficient. So the worst cannibalization is losing a sale.

GE was able to avoid the innovator’s dilemma regarding additive manufacturing (so far) because they kept an eye on what really mattered - making better product.

For much of the industrial sector, it’s hard to keep an eye on the main thing. What really do we mean by the term “better” product - for some customers, the cheapest / most efficient product is all they care about. But the reality is that if that is the answer, the OEM likely isn’t based in the United States, it has already been nearshored or offshored. Meaning the question many Midwestern Industrial companies need to be asking themselves about additive: what really matters to my customers beyond just making a cheaper part.

Playbook: Acquisition

In the Aughts and early Teens, GE built a lot of additive manufacturing internally. Allegedly they invested roughly $1.5 billion in manufacturing and additive technologies prior to any acquisitions. Then, in 2016, they acquired Arcam and Concept Laser, two European companies focused on this space, for a total of $1 billion combined. Prior to that, they acquired Morris Technologies and Rapid Quality Manufacturing in 2012. Morris Technologies was one of the earliest industrial metal-printing service bureaus in America and had deep expertise actually printing aerospace parts.

Arcam and Concept Lasers were about acquiring the right machines for the job. Morris and RQM were about know-how and production experience. They had to unlock that first before they even knew who to buy with a high level of confidence. Ultimately, that’s a talent-focused acquisition - buy people and process so you can buy the right tools later to back them up.

To illustrate that point, the LEAP Fuel Nozzle was actually developed prior to the Arcam and Concept Laser acquisitions. They had started figuring additive out, were impressed with the results, and started doubling down.

So you could say “sure, GE avoided the innovator’s dilemma, but there were market forces at play that drove them in that direction because of the demand for more fuel efficient engines”. But the problem with that logic is that GE started invested in additive manufacturing well before the tools were fully mature.

In fact, the acquisition that probably kickstarted all of this most effectively was that of Morris Technologies. That was the first additive acquisition that started driving things forward for GE’s additive practice. This led to the “first complex component made additively for production jet engine.”

Morris Technologies was founded by Greg Morris, Wendell Morris, and Bill Noack in the Cincinnati area, under the shadow of the GE Aviation Corporate HQ. GE started working with them in the 1990s. This helped GE better understand how additive manufacturing worked and helped Morris develop a reputation of one of the best job shops in the world. Morris is the company that produced the LEAP Fuel Nozzle. When that product was first delivered, people internally realized the power of additive and knew they needed to have them on the team.

Maybe GE doesn’t buy Morris if they aren’t up the road from them in Cincinnati. Or maybe Morris doesn’t get started if GE isn’t in their backyard. Maybe GE is still able to buy Arcam and Concept Laser without the Morris acquisition. These are sliding door moments that are hard to tell.

One thing is for certain: this didn’t just happen. Innovation sprung up from entrepreneurship and GE was smart enough to buy something that they needed. They participated in the innovation until it was obvious enough to buy the product. They didn’t bury their head in the sand.