A1) AM Disruption at Different Scale – What is Happening and Wild Guesses (2)

In my first two articles of AfT blog, I already showed you the MECE deltas as the key to AM disruption, why those MECE deltas with current AM economics are difficult working towards profitability and the vitality and concerns of extra gains from AM product (check this and this if you missed). I hope that my data evidence, derivations, and explanations have helped clear a little fog of AM hype and doubt. But, I know that most of our AM hope, hype, and “harm” lie in the future instead of the present. We used to believe AM would fundamentally revolutionize how we manufacture. Instead, now we tend to believe AM will be a complementary tool to how we manufacture. We can’t resist looking into future, making our projections and asking

How will AM technology be developed in next 5 years? What’s the trend for the next decade?

And will those MECE delta work in the near future?

So do I. In fact, making projections, or wild guesses have always been my favorite “sport” throughout my entire research life: technology development projections, technology adoption predictions, and sometimes even political election wild guesses. I always learn a lot when my prediction meets reality, from both the goodness of fit and deviation. Of course, I am not a fortune teller. I proceed as a strategic analyst and my “crystal ball” is data. For additive manufacturing technologies, I’ve been collecting data for years to make those projections. In this article, I will share some of my “wild guesses” about the vital deltas and AM disruption I discussed before, based on the data I found.

The first vital delta I discussed in my last article is the machine depreciation. With the current AM machine price, generally AM production faces a yearly cost of $20k-$120k (polymer) or $60k-$270k (metal) from the machine itself. The machine depreciation can represent a huge cost segment ranges from 40% to even 70%. Knowing how AM machine will be priced can greatly influence the landscape of many technology strategies, and of course, the delta profit map I brought up in the AfT. The difficulty behind this is that AM has many factors (e.g., size, speed, accuracy, service, brand) that may influence the price, and sometimes impossible to standardize under different operating mode and conditions. I’ve tried many factors to connect with machine price, but in the end, I only found one that may work. A very simple one: the volume of the building envelope/chamber, or in an approximate but simpler word, the functional size of the machine.

C1 Wild guess trend line

Based on the data that I was able to find for the AM systems in the past decade, you can clearly see a trend that machines with a larger building chamber (or in simpler word, larger machine) tend to be more expensive, and metal AM systems tends to be one order of magnitude more expensive than polymer AM systems with similar chamber volume.

You may think so what? This is common sense to many people who are not even familiar with AM industry. Here is the fun part –

If we take those data and analyze their trend line following the temporal sequence (e.g., biannually as showed in my following analysis and figures), the pattern of how those trend lines move in the plot is very interesting. The trait may demonstrate how the price of AM systems was driven down over the decade, and possibly suggest how we project the future.

C1 Wild guess polymer

The price of polymeric AM systems has been driven down significantly in the last decade and seems to continue dropping especially large machines. The significant price drop of newly released systems seems to start at 2013 with about 20% decrease and follow with an even larger drop (even close to 90%!) until now. Reasons can be the cheap machines with moderate performance got bigger, or expensive machines with great performance got cheaper. My guess is both. Due to increasing competitions, the profit margin of the polymer AM machine has been squeezed out quickly. We possibly see that the price of more and more systems is reaching a fair margin close to their cost, even this year – 2018.

C1 wild guess metal

Metal AM is a different story but may turn out to a similar ending as polymer AM. From 2007-2014, the price showed nearly no changes. Even there is, a 10% fluctuation biannually at most as shown by the data, which is probably within the range of error. The drop of metal machine price started in 2015 with a possible speed of 30% biannually, although I don’t have enough data to get a trend line for 2017 & 18. The decrease seems to show no preference for machine size. It makes sense since the metal machine with large building envelope tends to have extra powerful components, such as bi or quad-laser and in-situ monitoring. Those components can be very expensive, which raises the costs of building such systems significantly and possibly rule-out a larger space of price dropping for larger systems.

With over 250 data points, my dataset is definitely not exclusively, may not objective, or even not accurate enough to have a bulletproof prediction. But I will make my projections not too further way (for the year 2020) with my confidence anyway and call them “wild guesses”. I hope these “wild guesses” can help you regarding strategy for AM development and deployment.

My #1 wild guess: starting at 2020, all newly released polymer AM systems will cost no more than $150k, despite the performance, size or functionality.

(I am pretty confident about this one. Although with limited data, we can already see such trend line now 17-18. In fact, HP released new full-color 3D printing systems yesterday at Feb. 5th, 2018 with a price range from $50k to low $100k)

My #2 wild guess: the metal AM systems released in the year 2020 will cost 50% less compared to a similar size machine in 2015.

(Assuming a continuing 30% reduction as suggested in 15-16, it takes 4 years to reduce 50%. I even leave myself some space for error. 😊)

The second delta I discussed before is AM material costs, which lives the devil sometimes. The material price is extremely high (10-400 times more expensive) as I showed you before. The reason behind is complicated including the physics of the process, the supply & demand, and the business model. Producing AM materials with a desired quality and form sometimes require extra processing steps, which inevitably raise the price by some. Especially for metal AM powder, the requirement of spherical shape and particle size distributions limits the yield of the extra step of atomization, which chains the bottom line of the price (check this). Additionally, the AM material demand is still marginal comparing to traditional material markets (e.g., total metal AM material consumption is about 1000t), which the economy of scale is poorly leveraged. But more importantly, the high material price set by AM OEMs may not even reflect the material production costs. AM material can be the cash cow for many OEMs. You can easily buy similar AM materials for a much cheaper price from an independent material supplier, but at a risk of less optimum output and losing machine warranty. All these factors can be improved when AM markets grow, which drives down the material price, open up new markets and keep driving it down. So the price of AM material is predicted to drop in the near future

But by how much? Unfortunately, unlike the machine, the data for the material price is much harder to find or ask as a researcher. As a result, my projections on the material price will be closer to wild guess and based on my speculation and other people’s projections.

My #3 wild guess: By the year 2020, AM polymer material price will decrease by 60%

C1 wild guess carbon

(I made my speculation based on the recently released news by Carbon3D. Carbon plans to reduce their material price through a production-scale materials program, which aims at offering sub-$100 per liter material comparing current material price at about $250. The saving comes from the economy of production scale. Although we didn’t know how large their resin volume increase in production in the year 2017, but we can see with a similar range of volume increases, the price can hit a 60% reduction. I am betting such rule can be applied to other polymeric AM materials, and we can see such volume increase before 2020)

My #4 wild guess: By the year 2020, AM metal material price will drop by 40%

(Actually, you can buy 40% cheaper material now from independent suppliers. I am projecting that their material can reach the standard of the AM OEMs somehow by 2020)

Now I showed my projections for the year 2020 on the two major deltas I discussed in my last articles. Those new numbers for the year 2020 may change the landscape of AM disruptions. Such economics improvement can open up markets that you can charge extra $100/kg, but still may not be enough to tear up instead of knocking on the door of automotive industry or other industries with load bearing/engineering functional parts.

C1 wild guess delta

So my #5 wild guess: By the year 2020, AM will keep penetrating to the markets where deltas can be compensated by $100/kg, and we will not see a wild adoption in the automotive industry.

Here are my 5 wild guesses, do you agree or disagree?

Watch out for my next article for our blog “Additive for Thoughts” next week!



A1) AM Disruption at Different Scale – What is Happening and Wild Guesses (1)

In my last AfT blog, I discussed the importance of delta profit on AM disruption, and the MECE deltas that can be categorized as AM disruption on the process, product, and management. If the MECE deltas work toward a positive delta profit with realistic numbers of current technologies, we know AM can be disruptive.C1 equation

C1 Map edit

But are those MECE delta working?

In this blog, I will offer some rough estimates, rules of thumbs and run the numbers with you. You will clearly see the answer, understand what is happening of AM disruption, and why it is happening.

“Trust the process” – Joel Embiid, 2018 NBA All-star of 76ers

I will start with the scale of the process since this is where the disruption starts. Let’s say you want to add in AM capacity into your processing line today. The first thing is investing capital for machines and then immediately facing depreciation through ownership. Depending on the level of the machines, the depreciation plus maintenance would be about $15k-$120k/year for every polymeric AM machine and $60k-$270k/year for metal AM.
A1 table1

As for now, delta profit: deficit $15k-$120k/year (polymer AM), or $60k-$270k/year (metal AM) at the delta term of depreciation.

Printing parts directly sometimes can save mold, die, other tools, and the associated depreciation costs. This can be a more widely situation with polymeric processes. Take injection molding as an example, the depreciation costs saved of a mold can vary from $2.5k/year to $50k/year, depending on the complexity of the part and production volume. Considering this, the delta of depreciation under different conditions of part complexity and production volume:

With an AM machine sitting on your floor, you are facing ~$15k-$300k deficit from depreciation in most scenarios except if you plan to replace the production of complex plastic parts with small volume.

(Sounds familiar? This is one of the AM disruptions described as happening right now in many sources and materials. Examples include rapid prototyping and customized products.)
2018-01-29 (1)

Now let’s start printing. Every build consumes loads of materials, which combining with machine depreciation generally represents significant portions of AM total costs. But unfortunately, AM material is expensive. How expensive? Super, 10-400 times more expensive for most materials compares to its conventional form (bar, sheet or others). Even if you consider the AM advantage as a near-net-shape process with relative little scrap, plus the mass reduction generally achievable for most designs, the unit cost of material is still a raise for most polymers and metals, except expensive metals such as Ti64 or Inconel and a huge amount of machining involved for the original part.
A1 material

As a result, you will generally see a negative delta of the unit material of $10-40 per single polymer part, $10-70 per complex polymer part, and $50-400 per simple metal part (assuming 100g polymer part and 1kg metal part). You may see a negative delta of $50/part to a positive delta $600/part for complex metal applications depending on the level of complexity and material type.
2018-01-29 (2)

So in most scenarios, the delta profit will add another deficit from the unit material costs increase multiplied by production volume. Until now, AM production not only starts with a deficit but keeps adding deficit when you produce more. The only exception is complex parts with expensive metals to be machined away.

(Sounds familiar? This is one of the AM disruptions described as happening right now in many sources and materials. Examples include aerospace parts such as fuel nozzle. Counterintuitively, for this exception, you should not consider AM for small volume production as AM usually suits for. Instead, you should print as many as you can to cover up the delta deficit in depreciation. This may be one of the reasons for GE ramping up AM capacity quickly for the famous fuel nozzle production.)

A1 delta

Now you roughly see what is happening at the process scale with AM disruption. AM is facing extreme difficulties claiming a positive profit delta purely from the cost perspectives with current economics. I am not even counting on post-processing, inertia gas supply, and other possible cost increase. You may ask how about labor? I doubt AM production can save $50/polymer part (about 2 hours labor per part) or $200/metal part (about 8 hours per part) for most cases since direct labor generally represents a very small portion of total costs.

So my #1 core message of this blog: forget about the process scale if you consider current AM disruption for most cases. If it happens for certain applications or industries, the fundamental changes probably happen in the product or management. Think about other MECE delta terms in the map I offered you: Indirect Labor, S, G&A, unit price, unit volume….

Among all these terms, the most important delta you can count on to reverse the negative is the unit price. Whether it is lighter weight, better performance, or a longer lifetime, the unique features achieved through AM offer a competitive advantage in charging more for each part. It’s nearly impossible to squeeze $40/polymer part from direct labor or production supplies, but gain extra $40/part for faster delivery or better performance is achievable although may not be easy.A1 product

So my #2 core message here: product, product, product!

The current and near-term AM disruption would probably happen at the product-scale. The value added and the ability to charge extra from product innovation majorly determine the Top Line of the delta profit of AM. If you see a way to raise your product price with $80/polymer part or $500/metal part through AM, don’t hesitate to jump in and embrace the disruption.

But remember, raising price is always difficult. How can you prove and sell the value-added among all those uncertainties of AM processes and parts? The best case is the AM product is disruptive with nearly no competition, which offers you great power in asking whatever the price is if your customers can afford. Such case applies to nearly all major industries that AM currently penetrates deep, such as medical (customized crown and implant), prototyping (fast delivery), and tooling (informal cooling).

But sometimes the improvement is incremental, then selling the improvement may not be fun anymore especially when the product is a part of a system with hundreds of thousands of parts. This is due to the dilution of the improvement inside a complex system (assuming 20% performance improvement, and the part contributes 5% of the total performance of the system, after dilution the total improvement of the system from the innovative part is 1%). The assembler of those parts needs to consider finding a way selling single-digit improvement to final customers and balancing risks and uncertainties, which is difficult. Such scenario applies to industries such as aerospace and automotive, which you see a lot of integration of AM internally and pass vertically. Easier to sell, higher confidence. Also, you see GE is trying to build a turbine with as many AM parts as they can because it is easier to sell to the customer with a disruptive product adding up many single-digit improvements instead of a double-digit improved part.

Core message #3: disruptive final product (not part) directly to customers, if you think about best product strategy for AM technology.

All my discussions about the bottom line and top line of the delta are based on current or past AM performance. What about future? What’s the trend of AM development? How does the bottom/top line change in future? Keep your patience and I will give my predictions (or wild guesses) based past data I collected in my next blog in a few days. Please share, like, comment and follow our website if you enjoy this article!

My ending question for you:

What’s the exception successful case for AM disruption at the process, product and management scale? (I am also highly interested considering my personal limited scope)



Some of my references to those number:

  1. DS Thomas and SW Gilbert, Costs and cost effectiveness of additive manufacturing, NIST Special Publicaion, 2014
  2. M Baumers, Raw material pricing and additive manufacturing, EPSRC, 2014
  3. R Song and C Telenko, Material and energy loss due to human and machine error in commercial FDM printers, Journal of Cleaner Production, 2017