As many as ten million desktop 3D printers may be sold this year. This makes the desktop filament market potentially a significant business. I’ve interviewed key participants in this industry, users, and OEMs to develop a consolidated picture of the current market. I’ve made tables of major filament producers, compared pricing, and done analysis on the materials, prices, speed of delivery, marketing, resellers, and more. Then I used a stealth-mode deep-research AI business intelligence tool to index all pages of filament producers worldwide and all shop pages of 3D printer vendors. The tool indexed all forum mentions and blog posts, discussions, comments, and more about filament. This tool consolidated, analyzed, defined trends, and looked at the market as a whole. The research tool I’m using has capabilities way beyond publicly available LLMs, and delves much deeper into analysis and data. But this kind of capability is sure to become available more broadly.
So I did my analysis, listened to others to gather their views, and then separately looked to the AI tool for validation afterward. I would not recommend relying solely on indexing, crawling, or business intelligence software to do this. It’s only useful if you can spot errors, understand calculations, and have your own informed opinion. It’s also important to delve deeper when using any AI tool to understand why it makes certain assumptions and performs certain calculations. And in my opinion, it’s important to form your own opinion in advance in order to make the right query and understand if the tool is feeding you, and itself, the right information. I then sliced and diced the consolidated information in another tool, Google’s Notebook LM tool. I edited the information, removed some sources, and added others. Then I asked Notebook to generate a PowerPoint deck of some of the relevant information. This article was written entirely by me, and it demonstrates what someone can do with AI as a useful research tool. I’d never want AI to be my co-pilot; I want to land the plane, and I need to understand where and how we’re flying. Anyone who relies on AI to run your processes, intelligence, analysis, or judgment in the background without you understanding what you’re doing is an idiot. So I don’t want a co-pilot, nor do I want to cease doing critical thinking and run myself on autopilot. But I wouldn’t mind an AI Loyal Wingman. I mean, it’s not going to fly my plane, but it could take a bullet for me. As well as update you on the Desktop Filament Market today, this article will show you the ability of cutting-edge AI tools to aid and imperil research.
The Desktop Filament Market State of Play
Some of the initial filament players, low-cost Chinese brands, have entered into a killer competition with one another. eSUN seems to be weathering the storm somewhat, but other names have disappeared. Generally, we have smaller players, such as local 3D printing shops and national 3D printing-specific retailers, still shipping their white label filaments. In lockstep, the fortunes of Mitsubishi and Keene Village (now Toner) have waned, as they have had to secure new revenue sources and new ways to sell directly. The stores and retailers used to be a large market segment, but have stagnated and dwindled of late. One big player was initially responsible for this: Amazon, which brought in low-cost, semi-reliable filament. Then we have US and European premium brands such as colorFabb, which, through high quality and innovation, are trying to stay ahead. Other lower-priced brands, such as Spectrum in Europe and Protopasta and Atomic in the States, are pushing affordable but locally made filament with high-quality signifiers.
Prusa has a high-end, incredibly good brand and sells it to its community. Meanwhile, Elegoo (through SLA resin revenue), Creality (via the runaway success of High Speed PLA and Hyper), and Bambu (with ease of use) have shown incredible growth across the budget-to-premium segments. These players are now defining the market. BASF and the other polymer majors, meanwhile, are on the retreat, or have forgotten about us entirely by now. Stratasys has a portfolio of industrial filaments but does not seem to be engaging the desktop market. Evonik and Arkema are also going the specialty route. Another form of specialty is flexible, in which Recreus and NinjaTek have strong brands, and no one has really managed to make amazing TPU work well. 3D4Makers and others, meanwhile, sell high-end filaments for space and defense.
The above is a Notebook-generated slide based on my scraping data. Now, I agree with the segmentation. But. I completely disagree with the “completely different buyers” part. Sometimes, people, companies, and schools buy cheap filament. Some people use inexpensive draft filament most of the time, but buy expensive engineering materials as well. Some people use draft filament and then buy premium filament for nice things that they will keep for a long time. Also, many people and companies buy filament from multiple sources. I may purchase material from a local store, but will buy from 3D4Makers if I need PCL, from colorFabb if I need foaming, from Spectrum if I need a lot of something, and sometimes from Bambu if I have a great deal. Meanwhile, most of what I use every day is Prusament. So I think that the assertion at the bottom is wrong. What’s more, I have looked through my data, and I have no idea where the LLM got this from. I don’t have a single piece of information that implies this is true, that the filament is so elastic, or that buyers are so consistent. Also, I do agree that the segmentation is nice, but there are lots of firms that buy aerospace materials and cheap PETG. And there are lots of firms that buy PEEK and other premium materials such as PC. So I think that the segmentation does not at all “obscure five separate markets.” So, this nicely presented slide has some order to it, but could be completely misleading.
Market Segmentation
The dragons have given us many splendored colors and rainbow filaments. We’re getting more “safer” filaments, recycled ones, and ones fit for purpose, such as rail-certified materials with Fr. Companies like Tectonic 3D are engineering incredibly high-performance materials designed to work on desktop systems. We’re seeing more non-NatureWorks PLA and lower-cost feedstocks emerge, as people turn to commodities to keep margins alive. Some are working more with larger compounders, while still others are vertically integrating more. Chinese players are expanding their distribution and manufacturing operations in (or to) Europe due to tariff concerns. Many European players are turning towards defense applications. PLA Plus and High Speed PLA are profitable but often a mirage, not really delivering meaningful performance in many cases. Filled materials will become more prevalent, and interesting eco-filled materials will do well. Whether the algae, coffee, or whatever actually conveys environmental benefits remains to be seen.
Print farms are becoming very important accounts for some filament companies. Some farms already make their own filament. Others get good deals from the big Chinese players. More should work with premium suppliers to unlock ongoing contracts at low costs. Quality has risen precipitously across the board. But many low-cost brands still persist with tolerance issues.
Fault Modes
Under-dimensioning filament to reduce clogs still occurs, and many brands still have issues with ovality. Filament dryers and user-made dry boxes, as well as commercial variants, are making life easier for users. Many users have become significantly better at storing filament. Packaging is also improving. There are better bags for storing your filament now, and spools are becoming more sustainable.
Breakage and nozzle clogs are still issues. In AMS-like units, grooving or feeders digging into the filament are also problems. Higher-speed machines have caused some filaments to perform much worse than others. One of the biggest user fault modes now is winding and tangling issues coming about through user error (unpacking and feeding while pushing filament under another wind) and bad winding coupled with higher speeds. Filament clips and correct user unspooling and insertion were nice to have, but currently are important. Adhesion issues have subsided for many users, but nozzle degradation over time and general filament degradation are still issues.
As you can see above, the Bambu store visits are very high when compared to pure play filament brands. Bambu has caused a wave of people to, in fact, be better at printing and working with other people’s filament. There is also a real convenience lock-in. Bambu and Creality’s apps and overall software experience increase prints and help drive filament sales. I don’t think that the AMS provides a real lock-in or razor-and-blades model. I think it’s smarter than this. I’ve tried many different novel filaments, both on the AMS and loose, on several Bambu systems, and the printers work very well with them. In many cases, the specific settings can be entered very easily. Instead, they make it convenient. And some users, especially new ones and neophytes, love this. Traditionally, in the filament market, we saw that new users remained with the printer-supplied filament for months before venturing out into new brands. What Bambu is doing is collecting an awful lot of user-specific usage information, filament information on aggregate, and trends. A firm that can monitor downloads to the app to the camera can judge different filaments and see what power users are turning to. They can then white-label the right filaments, sell the colors that work, develop missing colors, and focus on grades that perform poorly. Its more of a gap analysis masterstroke than razor-and-blades to me. Its not that you have to use the Gillette razor for it to work. It’s as if the Gillette razor could scan everyone’s face and then develop blades specifically for your skin and face type. This is the true power of the total control business model. It’s not a razor blade, it’s a facehugger.


























