Bambu Lab has not announced an H3C, and nothing in this post comes from Bambu. What has happened is that a pile of regulatory filings, patents and forum chatter has built up around the name over the last two months, and it is worth getting the picture straight early, before launch-day noise buries it. This is a starter post. It will be wrong in places by the time the machine is real, and the aim is to be clear about which parts are solid, which are reported, and which are hope. The R1 teaser post went through the same cycle, and the lesson from it was that a teaser confirms less than people assume.
How to read this post
| Claim | Status | Basis |
|---|---|---|
| A Bambu Lab printer labelled “H3C” exists | Confirmed | FCC filing, ID 2A6J8-PF008P, dated 11 September 2026, model PF008-P |
| A “P3C” printer exists | Confirmed | FCC filing, ID 2A6J8-PF006P, dated 15 September 2026 |
| An “X2D Pro” exists | Confirmed | FCC filing reported the same day as the H3C |
| AMS 3H and AMS 3P exist | Confirmed | FCC filings dated 21 September 2026 (SA017 and SA018) |
| A handheld “S1” 3D scanner exists | Confirmed | FCC label and label-location drawing |
| Bambu is working on a ten-nozzle hotend-changing printer | Confirmed | Three Chinese patent applications published 14 August 2026 |
| The H3C is that ten-nozzle printer | Speculation | No document links the two, and Bambu has not said what the letters mean |
| What the AMS 3H and 3P do, and which printers they suit | Unknown | Labels only, with photos and manuals confidential until March 2027 |
| Any “leaked photo” of the H3C | Unverified | Disputed forum images, see below |
What is actually confirmed: the FCC filings
The solid ground here is regulatory paperwork. Shenzhen Tuozhu Technology, the legal entity behind Bambu Lab, registered a printer with the US Federal Communications Commission on 11 September 2026. The label identifies it as an H3C, model number PF008-P. The filing shows a rear power inlet and dual-band Wi-Fi, and an X2D Pro filing from Bambu appeared in the same database the same day. Four days later a P3C (model PF006-P) was certified, again by the same company. Neither name has been explained by Bambu.
It is worth being precise about what an FCC filing is and is not. It proves a device exists and has been certified for radio emissions. It does not reveal a specification, a price or a date. Bambu also asked the FCC to keep the external photos, internal photos, user manual and test-setup images confidential, which is a standard short-term option. That matters for the “leaked photos” question below, because it means no photo of the machine came out of the FCC filing.
One timing pattern gets quoted a lot. Bambu’s recent printers have shown up at the FCC within days of launching: the H2C was certified on its launch day and the X2D two days before its launch. By that logic an H3C announcement might be close, but the H3C filing is now a month old and nothing has appeared, so the pattern is a hint rather than a schedule. Confidentiality on the photos and manuals runs until March 2027, which has been picked up as a possible outer limit, but it is a standard confidentiality window rather than a launch date.
The patents: where the ten nozzles come from
Separately from the FCC filings, China’s patent office (CNIPA) published three Bambu Lab applications on 14 August 2026. They describe a printer with ten nozzles: nine interchangeable printheads stored in a rack, plus a tenth on the toolhead itself, matching ten filament spools drawn on a spool rack in the patent figures. For comparison, the H2C’s Vortek system stores six hotends plus one fixed liftable nozzle, so this is a substantially bigger magazine. The Vortek post covers how the current swap works.
A few details from the patent write-ups are genuinely interesting engineering rather than filler:
- Tightly packed docks. The spacing between adjacent dock positions is smaller than the toolhead’s own width along the travel direction, so more printheads fit in the same footprint.
- Unlock and dock in one motion. The printhead’s travel from engaged to released position is its travel into the rack, cutting a step from every swap and removing any moment where the head is loose.
- Three locating members per dock, each removing movement along one axis, plus an angled magnetic seat that pulls weakly on approach and strongly once seated.
- The bed lowers first when a swap happens at a print height between 0 and 20mm, because parked printheads hang lower than the printing nozzle and would otherwise clip the bed or wiper zone early in a print.
- A light inside the cooling duct. A separate toolhead application describes illumination routed through the part-cooling duct, with matte black and matte white inner surfaces, so a camera sees a lit nozzle above a dim print area.
Two cautions apply. Patents describe mechanisms and never name products, and companies patent things they never ship, so these documents say nothing about timing. And the early publication is notable: the applications went from filing on 30 June to publication on 14 August, about six weeks rather than the usual 18 months, which only happens when the applicant requests it. Makers101’s patent analysis, which is the most thorough write-up, was published on 21 August, before the FCC filings, and says the H3C is only a community name. That was a fair reading then. The FCC label has since put the name on a real regulatory document, though what the letters stand for is still unexplained.
Whether the patented printer and the H3C are the same machine is speculation. The case for it is circumstantial: the H3C filing appeared shortly after the patents were published, and the H-series has become Bambu’s platform for more advanced technology. The case against is that nothing links the two directly. Some of the nozzle-counting coverage, such as “a ten-hotend toolchanger”, also overstates it, since the patents describe nine in the rack plus one on the toolhead, which is a little different from ten fully swappable tools.
The “leaked photos”: handle with care
This is the part that deserves the most scepticism. Three different things get lumped together as “leaks”, and only one is solid.
The FCC label images (solid, but dull). The public part of the filings is regulatory labels and label-location drawings, not product photos. A forum thread from 14 September started from a screenshot of one of these. Members read what they could from it: the label shows the rating as 8A at 110-120V AC, which one member noted is lower than the H2C’s roughly 1,250W, and the markings include Mexican and Brazilian compliance symbols with no EU marking. One member concluded it was destined for Latin America. Another pointed out that the FCC regulates emissions regardless of supply voltage, and that Mexico and parts of Central and northern South America use the same electrical system as the US, so a single label is thin evidence of market.
The power-rating speculation (interesting, unproven). The lower rating prompted a lively thread of guesses, all unverified. One suggested the machine might use zoned bed heating like the A2L, another that it might be a more efficient heating strategy, and one even raised the possibility of a change to inductive heating, with the comment that if induction were going away, the current Vortek design would be “toast”. Another member claimed the published patents say the feeder will not be actively heated and will use recirculated chamber air to dry filament, which could also explain lower power use. I could not verify that claim against the patent text, so it should be treated as a forum comment rather than a finding.
The “leaked image” of 7 October (unverified and disputed). A forum user posted an image labelled as a leak of the machine. Another member replied that an AI detector rated it 94% AI-generated and that it contained obvious mistakes, particularly at the bottom. The poster responded that the image appeared in two videos from reputable sources, and that AI had been used to remove the background and reflections to protect the leaker. Beyond the specific dispute, there is a structural reason for caution: because Bambu asked the FCC to keep the external and internal photos confidential, any picture of the H3C circulating now cannot have come from the FCC filing. It would have to come from somewhere else, and an image that has been through AI background removal has already lost some of its evidential value.
AMS 3H and AMS 3P
The rumoured new AMS is real, in the narrow sense that two new Automatic Material System models have cleared the FCC. They were certified on 21 September 2026, the day before the R1 laser launched and a day before the S1 scanner filing, and they sit in the database between the two.
| AMS 3H | AMS 3P | |
|---|---|---|
| FCC ID | 2A6J8-SA017 | 2A6J8-SA018 |
| Model | SA017 | SA018 |
| Grant date | 21 September 2026 | 21 September 2026 |
| Power input on the label | A mains-style input, which points towards heating | 24V at 1A, a quarter of the AMS 2 Pro’s rating |
| Photos and manual | Confidential until 20 March 2027 | Confidential until 20 March 2027 |
What the filings do not reveal is long: how many spools either unit holds, whether the 3H actually dries filament and at what temperature, which printers each unit works with, the price, and the release date. The reasonable inference from the labels is that Bambu is preparing two different units, one with a mains input that suggests heating and one that is low-power and probably unheated.
Here is my own guess, clearly labelled as one: the “H” and “P” in the names mirror the printer series, with the 3H aimed at H-series machines and the 3P at the P series, much as the AMS 2 Pro was introduced with the H2D. Nothing in the filings says so. If it is right, the lower-power 3P would be the unit for the more affordable machines, which would fit the speculation that the P3C is a value-oriented follow-on to the P2S. For readers like me running the A1 and A2L, the practical question is whether any of this touches the AMS Lite, and for now there is no evidence it does. The top mount post and the dual BCMU verdict are the relevant background on how the AMS Lite setup behaves today.
The rest of the September filing spree
The H3C sits inside a broader cluster, which is part of why the hobby has been so excited. In the space of about two weeks the FCC records revealed an unannounced printer, a second unannounced printer, a printer variant, two AMS units and a handheld 3D scanner. The scanner, called the S1, is only known from a public label and a label-location drawing showing a body with a hand grip, which suggests handheld, and from nothing else. It ties in with the 3D scanners post, where Bambu was notably absent from the field, and with the broader pattern of Bambu pushing into adjacent categories covered in the R1 launch post and the UV printer post.
Timeline so far
| Date | Event |
|---|---|
| 6 November 2025 | Priority date claimed by the placement-rack patent applications |
| 30 June 2026 | Three patent applications filed in China |
| 14 August 2026 | The three applications published, after an early-publication request |
| 11 September 2026 | H3C and X2D Pro appear in FCC filings |
| 14 September 2026 | Forum thread on the H3C label begins |
| 15 September 2026 | P3C certified by the FCC |
| 21 September 2026 | AMS 3H and AMS 3P certified |
| 22 September 2026 | S1 scanner filing, and the R1 laser launches |
| 7 October 2026 | Disputed “leaked image” posted on the forum |
| 20 March 2027 | Confidentiality on AMS photos and manuals ends, with similar windows for the printers |
The hope and speculation section
Everything from here is wish-listing, so treat it as such. These are the things the community is hoping for, and the ones that would matter most for the kind of printing this site covers.
- Far less purge waste. The hidden workflow cost post showed how heavily purge waste dominates complex multicolour jobs on a single nozzle. A system that keeps many materials in separate nozzles attacks that directly, which is the same reasoning that makes the Sovol M1D and the LightMake L4 interesting.
- A purge-tower-free path. The Stratasys verdict post covered a patent fight over the purge tower and the legal pressure on Bambu’s current approach. A hotend-swapping design that needs far less purging would also change that picture, though I would not read legal strategy into the hardware without evidence.
- Reliability over headline count. The next big shift post argued that dual nozzles, reliability and ecosystem matter more than raw speed. Ten nozzles is a spec-sheet number, and the real question is whether the swap mechanism stays reliable over thousands of changes.
- A sensible AMS pairing. If the 3H does actively dry filament and the 3P does not, a sensible pairing across the range would be welcome, especially if the 3P brings active airflow drying logic down to cheaper printers.
- Sensible pricing. The R1 landed at $2,499, undercutting comparable lasers and, as one review noted, even Bambu’s own P2S. If the H3C follows the H2C and X2D pattern, it will sit at the top of the range, but a P3C could be the more affordable story.
- A better answer to Bambu’s competitors. One write-up quotes community talk that Bambu is “behind on toolchangers” relative to Prusa’s XL and INDX and Snapmaker’s U1, so an H3C would be strategically important if it ships.
Should this change what you buy?
One forum member said they would hold off on an H2C purchase until the H3C is confirmed, which is a sensible instinct taken a step too far. There is no date, no price and no specification, so waiting is a bet on an unknown release window. The broader advice from other coverage is to buy on what is in stock for a machine you need now, and nothing in these filings changes that. If you are weighing an H2C or X2D purchase for work you have waiting, the filings are a reason to keep an eye on Bambu’s announcements, not a reason to put your plans on hold.
What to watch for
- An official Bambu teaser or announcement, which will explain the letters if nothing else.
- Further FCC exhibits becoming public, though the photos and manuals stay confidential into 2027.
- Any retailer listing, regional certification or customs record naming the models, which often surface shortly before launch.
- A Bambu Studio or firmware release adding profiles for an unannounced machine, which has been an early giveaway for earlier launches.
- Patent publications that tie the rack design more directly to a named product.
The honest summary
Something called the H3C exists, and so do a P3C, an X2D Pro, two new AMS units and a handheld scanner, all certified within about a fortnight. Separately, Bambu has patented a printer that swaps between nine docked printheads and a tenth on the toolhead. That is the full list of things that are confirmed. Everything connecting those two facts, the machine’s speed, price, size, heating method, spool count and release date, is guesswork, and the circulating “leaked photo” should be treated as unproven at best. It is still a genuinely interesting set of clues, and it points the same way as the rest of the industry: towards more nozzles, less waste and more intelligence in the filament system. This post will be updated as soon as Bambu says something on the record.
