AMPulse

AM funding events 2021-2026, one instrument

The Western Winter

The AM funding pullback since 2022 is told as one industry-wide winter. Counted inside a single instrument, the winter has a geography: Germany's equity count falls 40% and its supply-side count halves, US equity events fall 39%, and 2025 is China's record year. The record is written on smaller cheques, and a quarter of it is state money.

AMPULSE DATA REPORT · 2026Data snapshot: Sep 7, 2026

Three weathers

The AM funding story since 2022 has been told at the level of the whole sector. AM Ventures opened 2024 calling 2023 "a less rosy AM year marked by venture capital (VC) downturn," with $1.1B invested "in the AM sector"; the unit of the story is the sector, and its exhibits are Western: Desktop Metal from a SPAC listing that valued it near $2.5B to Chapter 11 in 2025, Shapeways into Chapter 7, and a stretch of mid-2026 in which trade press counted 3DEO and Wurth Additive closing while BigRep's listed parent headed into liquidation with its operating business sold. Nothing in that telling is false about the total. What the sector-level unit hides is geography.

Counted country by country, the one winter is three weathers. Across 3,558 funding events in the instrument, 2,012 attach to Chinese, American or German companies, and those three series disagree. Germany logged 64 events in 2023 and 37 in 2025, 31% below its 2021 count. The United States logged 240 in 2021 and 189 in 2025. China logged 144 in 2025, its highest count in the series, 30% above 2021, spread across 117 companies, also its record.

The equity cut

Funding events are not all the same kind of money. Strip the series to dilutive stages, pre-seed through growth, and the three weathers separate further.

US equity events fell from 116 in 2021 to 71 in 2025, down 39%, while the US non-equity layer held, 124 events in 2021 and 118 in 2025, grants its largest line in both years (46, then 46) with strategic capital and debt growing inside it. The American winter in this data is specifically an equity winter, cushioned by the layer that kept flowing. Germany's is unconditional: 12 equity events in all of 2025, against 18 in 2022 and 19 in 2023. And China's does not appear: 91 equity events in 2025, the record for the series, up from 61 at the 2023 trough.

What the money weighs

A count record can sit on top of a dollar drought. The disclosed-amount record carries cumulative-total misreads, but medians do not care about a few misread outliers, so the question can be put to the parsed subset: the equity events whose round size the record holds in dollars. That subset is 77.5% of US equity events in 2025 and 75% of German ones. It is 30% of Chinese ones, because Chinese rounds are announced as "hundreds of millions of yuan" far more often than as a figure, and that gap is the reason every Chinese dollar figure below is read as a direction and not a level.

The drought is real, and it is everywhere. The median disclosed equity round fell from $19.3M to $8M in the United States, from $8.3M to $5.8M in Germany, and from $15M to $10M in China. China's record year is a record of count, written on cheques 33% smaller than the ones its 2021 got. So the geography is one of deal flow, not of capital: the West stopped doing rounds, China kept doing them and did them smaller.

The big cheques did not disappear; they moved. Split every country's equity events by whether the funded company's target market is aerospace and defense, and the $100M rounds of 2025 sit there: 5 of the United States' 9, 1 of Germany's 2. China's largest round of the year, Galactic Energy's $333.3M series D for a rocket maker, carries no target-market label in the census and so sits outside its aerospace column, which is why that column shows 0 rounds at $100M; read with it, the pattern is the same in all three countries. The median aerospace round in 2025 was $15M in the United States over 21 disclosed rounds and $42M in China over 4, against $7.1M and $7M for everything else. The companies raising those rounds, rocket and hypersonic makers, a fusion startup, two defense-drone companies, are in the census because additive is inside their production, not because they sell it. What is left after them is the AM supply side, and that is where the winter is coldest.

The supply side alone

Narrow to the narrowest cut this record allows: equity events of companies whose census label is hardware, materials, software or post-processing equipment, outside the aerospace and defense market. It is a label cut, and the labels leak both ways. Radiant (microreactors), Proxima Fusion (a fusion device) and Lyten (lithium-sulfur batteries) carry a Hardware or Materials label with an Industrial market and stay inside it; Divergent, whose product is an additive production system, carries a Platform label and falls outside. The cut is the closest the labels come to the AM supply side, not the supply side itself, and the $100M rounds it holds are named below.

US supply-side equity events fell from 53 in 2021 to 30 in 2025, down 43%. Germany's fell from 12 to 6. China's went from 47 to 56, against 56 in 2022, with a median disclosed round of $7M. On the cut that excludes every rocket company and every defense contractor, the three weathers are still three: the American AM supply side lost 43% of its equity deal flow, the German one 50%, and the Chinese one is at its high.

What China's record year is made of

The record is not a seed bubble, and it is not a desktop-printer story. Of China's 91 equity events in 2025, 40 sat at series A and 23 at series B. The 77 companies that raised them break by target market as 56.6% industrial, 22.4% aerospace and defense and 11.8% consumer, so ELEGOO, with a series B among its 3 recorded events, and Snapmaker, with a series B of its own, are the visible edge of a record that is mostly industrial. The pull behind that industrial share has names: BLT tied its FY2025 rebound, revenue up 40%, to deeper aerospace penetration and new consumer-electronics applications, and Apple's 2025 titanium Watch cases put consumer electronics on the list of production categories a Chinese metal-system maker can raise against. Widen to every channel and 117 Chinese companies were funded in 2025: 62 make hardware and 20 materials, 7 are service providers, and their median founding year is 2018.

Two more properties separate the Chinese record from a bubble. 40 of the 77 companies were raising equity for the first time in this record and 37 had raised before, so the year is half new entrants and half a pipeline moving up. And the pipeline is the one that graduates.

28.9% of China's 2021-22 seed-to-A cohort has reached series B or later, against 20% of the American cohort and 10.3% of the German one. Germany's number is the winter in one figure: of 29 companies that raised early money in 2021-22, 9 have raised equity again at all.

The third property is who wrote the cheques. Read the investor lists of the 77 Chinese companies funded in 2025 and mark the ones naming a government guidance fund, a state-owned group, a national industry fund or a development-zone vehicle: 21 of them do, 27%. The United States' 2025 list has 2 such companies. The Chinese leads are explicit about it. Vilory's series C was led by a national-level industry fund, Hongzhen Aerospace's angel round by a national fund with central, provincial and municipal guidance money behind it, Deep Blue Aerospace's series B by the Taishan Industrial Development Investment Group. The count is a name heuristic over free text and undercounts state money routed through nominally private managers; it is a floor, and the floor is a quarter of the record year.

The largest rounds of 2025

The six largest disclosed equity rounds per country, with the lead investor where the record names one. They are the rounds a reader will recognise, and they say where the big money in "AM funding" went.

Largest equity financings of 2025 by parsed dollar amount, per country, one row per company and stage. Amounts converted at the event-date rate; a yuan or euro round is shown in dollars. Event dates are month precision, and the lead is the one the record's sources name: ELEGOO's January 2025 series B and its DJI lead are as TechNode and Fabbaloo coverage recorded them.

Nothing at the top of any list is an AM supplier. Helsing, Castelion, Impulse Space, Boom, Galactic Energy and Deep Blue are in this census because additive is inside what they build, and their rounds are defense and space rounds. The AM suppliers in the panel are ELEGOO, Vilory and Avimetal, with 3D Spark and UniteLabs as software; Lyten makes batteries, Radiant microreactors and Proxima a fusion device, all three inside the supply-side cut by label, and Divergent is the one company on the three lists whose product is an additive production system. A reader who wants the size of the AM industry's own funding should read the supply-side chart, not this panel. The panel is here because these are the numbers the trade press quotes as AM funding.

The counter-thesis

The strongest case against this reading is that the instrument sees China better in 2025 than it saw China in 2021. The sweep ran in one window, so no year was collected before another; but if the sources it read cover 2025 Chinese rounds more completely than 2021 ones, part of China's rise is an undercount of its earlier years, and nothing here can rule that out. What survives is that the rise holds on every cut that would expose one: on equity events, on the supply-side cut, on distinct companies rather than events, and on financings after the same-round collapse. A source that got better at seeing China would have to have got better at seeing all of those at once. Closures are the same objection in another form, and they do not touch a count: a Chinese company that stopped raising keeps the events it had, so an undercount of Chinese deaths inflates the active cohort and moves no series.

The second objection is that a count record is not a funding record. It is not. The median Chinese round is 33% smaller than in 2021, and the subtitle says so. The claim that survives is narrower and, for an operator, more useful: Chinese AM companies are still closing rounds at a record rate, on smaller cheques, with the state as a quarter of the table. A German AM company had a 10.3% chance of reaching series B from its 2021-22 seed; a Chinese one 28.9%.

What each reader should take

What to watch

• 2026 full year. Through the snapshot the year-to-date counts run China 75, the United States 111, Germany 10; events reach the record with a lag, so these are floors. Where they land by the next sweep says whether Germany's 2025 was a trough or a step down.
• The Chinese median. $10M in 2025 on a 30% parsed share. If the parsed share rises and the median holds, the count record becomes a funding record; if the median keeps falling, it is a fragmentation.
• Graduation of the 2023-24 cohorts. The 2021-22 cohort's graduation rate, 28.9% China, 20% United States, 10.3% Germany, is the baseline; the same cut on the 2023-24 seed-to-A cohorts, run in 2028, tells whether the German number was the winter or the country.
• Section 849. From 2026-12-18 covered AM machines from covered AM companies are out of US defense procurement and operation. Whether Chinese equipment makers' 2026 rounds get larger or fewer after that date is a test of how much of the Chinese record depended on US defense customers.
• Aerospace share of the $100M rounds. 5 of 9 in the United States in 2025, and the rest went to a microreactor, a battery and a production-system company. A year in which a machine or feedstock maker closes a $100M round is the year the supply-side winter is over.

Lee, S. (2026). The Western Winter. AMPulse Data Reports. https://www.ampulse.online/reports/western-winter-2026