Wavester retweeted
Seems like the blue pen speculation was spot on. $APRTEC
$APRTEC
Image from BEX in Rimini, taken at the Zoppas stand.
Key people in the image include:
Lars Almhem, CEO APR Technologies
Robert Thorlund, VP Technology APR
Federico Zoppas, Zoppas Industries
Michele Peterle, Business Unit Director Aerospace & Medical, Zoppas
Note the table: two sheets of paper and two pens, neatly placed. Looks very much like something was signed, in duplicate. It would be odd to stage a scene like that with the Zoppas owner in the middle otherwise.
APR guided for a framework agreement with Zoppas in Q3. Q3 ends tomorrow. Perhaps we get a last day announcement after all and the target is hit.
Seems very promising. $APRTEC
Wavester retweeted
Had a very insightful call yesterday with APR Technologies CEO Lars Almhem, founder and former CEO Peter Nilsson, and AAIG with @Mark_AAIG and @MarkosAAIG .
Here is roughly where the $APRTEC thesis stands today.
APRTEC is a small Swedish team tackling liquid cooling problems with its EHD technology.
The main focus increasingly appears to be data center cooling, putting them in roughly the same arena as Accelsius, $INV and ZutaCore.
But the technology reaches much further.
They already sell space cooling applications and appear close to signing a commercial agreement with Zoppas.
They also have battery cooling solutions using immersion cooling, for applications such as drone batteries.
Beyond that, they clearly have their eyes on areas such as HBM, photonics and lasers, although the limiting factor today seems to be resources rather than opportunity.
The company has a very promising technical team and R&D background.
The main question is now simple:
Can they scale manufacturing and actually sell the technology?
If they can, I think the story has genuine 20x+ potential.
And some of the newest datapoints are starting to make that scenario more tangible.
When asked about factory expansion, management said they already have preliminary plans for a second and third factory, likely outside Sweden.
One factory represents roughly 500k units of annual capacity.
Blended unit pricing is now expected to be around $75 to $100.
That means the manufacturing footprint they are already thinking about could eventually support a very substantial revenue base.
Peter also said that the volume ramp is going very well so far.
We also know APR is in discussions with several large technology customers and hyperscalers.
One customer has already placed a pre production order, which has been delivered, and the product is now in field testing.
It is still difficult to know the exact timeline or whether this turns into a large volume order.
But what we do know is that a very large customer has already spent meaningful time and resources evaluating the technology.
Given that volume production is expected to begin in 2027, I would not be surprised if we get much more clarity over the next quarter or two.
Another interesting development is retrofit.
Historically APR has focused much more heavily on getting designed into new systems early in the development process.
They are now actively looking at retrofit opportunities as well.
Definitely something to watch.
The team is also expanding across operations, quality and sales, exactly the functions you would expect them to strengthen ahead of a manufacturing ramp.
And importantly, management has now confirmed gross margins above 50%.
That matters enormously for the long term valuation framework.
So the thesis remains very asymmetric.
The technology looks increasingly credible.
The commercial pipeline appears real.
Manufacturing capacity is being planned beyond the first factory.
Margins look excellent.
Now they need to prove that qualification turns into real volume.
I’ll publish more of the research later on my Substack, and we will likely also see deeper work from AAIG on their platform.
I’ll also update my original deep dive with the newest information.
For now I’ll link the older piece below.
And if you are interested in liquid cooling more broadly, definitely check out AAIG’s work on the other names in the space. $INV $APRTEC.ST $APRTEC $NBIS $VRT $DELL
@Mark_AAIG me and @emilhartela just had a great call with APR Technologies as part of our broader liquid-cooling sector research, speaking with CEO Lars Almhem and co-founder and Chairman Peter Nilsson.
We discussed their Rheion technology, which uses electric fields to move dielectric coolant directly at the chip level. They are starting to scale including customer-funded development work with a major global technology company. Credits to @emilhartela for bringing this company on our radar and make sure to check his Substack write up on it!
$APRTEC.ST $VRT $NBIS $DELL $INV
Wavester retweeted
$APRTEC valuation math for those interested.
Because APRTEC has minimal coverage and a huge amount of uncertainty, I figured I would lay out roughly what kind of revenue and valuation framework one can build from the information we have today.
We know they are currently building Factory 1.
Capacity: up to 500k units per year
Unit price: roughly $75 to $100
Gross margin: 50%+
Capex: relatively low, with future factories expected to be materially cheaper
We also now know they already have preliminary plans for Factory 2 and Factory 3, likely outside Sweden.
That gives us at least something to work with.
Factory 1 is expected to start ramping volume production in 2027.
The 300k unit figure has been mentioned several times, so my guess is that this is roughly the kind of annualized volume they are targeting during 2027.
But that could easily be an exit rate rather than full year production.
So let’s be more conservative and call FY27 production 200k units.
At $75 to $100 per unit, that gives:
$15m to $20m revenue.
With 50%+ gross margins, gross profit would already be meaningful, although after hiring, factory ramp costs, R&D and everything else, actual net profit would likely still only be a few million dollars.
2028 is where the math becomes much more interesting.
If Factory 1 ramps successfully during 2027, it is conceivable that Factory 2 and Factory 3 could start contributing during 2028.
Three factories at full capacity would represent roughly:
1.5 million units.
At $75 to $100 per unit:
$112m to $150m revenue.
If the business eventually manages something like a 20% net margin, that would translate into roughly:
$22m to $30m in net profit.
This is very much a scenario where execution goes right.
But importantly, it does NOT assume that a major hyperscaler or other customer finances a much more aggressive expansion.
Beyond 2028, modelling becomes almost pointless.
Cooling demand should accelerate massively toward 2030, and Peter recently mentioned that one customer had asked about something in the range of 5 million units.
APR obviously cannot deliver anything close to that today.
But if a large customer genuinely wants the technology, capacity could probably be expanded much faster than the path outlined above.
Now jump back to the 2028 scenario.
$30m of net income is roughly SEK 300m.
APRTEC today trades around SEK 800m.
If they actually reach that level of profitability while still growing rapidly in one of the hottest infrastructure markets in the world, I seriously doubt the market gives them a normal industrial multiple.
At 50x earnings:
SEK 300m × 50 = SEK 15bn valuation.
That is close to 20x today's market cap.
Obviously this is the bull case.
$APRTEC is still a scaleup. There are numerous hurdles between here and there.
Customer qualification can fail. Orders can be delayed. Manufacturing can be harder than expected. They may need additional financing. There will probably be some dilution. And in the true downside case, they simply fail to commercialize the technology and eventually run out of money.
But that is what makes the setup interesting.
Worst case: the scaleup fails.
Middle case: you own a tiny company with an engineering team that has apparently solved problems even NASA struggled with.
Best case: successful volume commercialization creates something approaching a 20x+ outcome.
The range of outcomes is enormous.
NFA. DYOR.
Deep dive below.
Had a very insightful call yesterday with APR Technologies CEO Lars Almhem, founder and former CEO Peter Nilsson, and AAIG with @Mark_AAIG and @MarkosAAIG .
Here is roughly where the $APRTEC thesis stands today.
APRTEC is a small Swedish team tackling liquid cooling problems with its EHD technology.
The main focus increasingly appears to be data center cooling, putting them in roughly the same arena as Accelsius, $INV and ZutaCore.
But the technology reaches much further.
They already sell space cooling applications and appear close to signing a commercial agreement with Zoppas.
They also have battery cooling solutions using immersion cooling, for applications such as drone batteries.
Beyond that, they clearly have their eyes on areas such as HBM, photonics and lasers, although the limiting factor today seems to be resources rather than opportunity.
The company has a very promising technical team and R&D background.
The main question is now simple:
Can they scale manufacturing and actually sell the technology?
If they can, I think the story has genuine 20x+ potential.
And some of the newest datapoints are starting to make that scenario more tangible.
When asked about factory expansion, management said they already have preliminary plans for a second and third factory, likely outside Sweden.
One factory represents roughly 500k units of annual capacity.
Blended unit pricing is now expected to be around $75 to $100.
That means the manufacturing footprint they are already thinking about could eventually support a very substantial revenue base.
Peter also said that the volume ramp is going very well so far.
We also know APR is in discussions with several large technology customers and hyperscalers.
One customer has already placed a pre production order, which has been delivered, and the product is now in field testing.
It is still difficult to know the exact timeline or whether this turns into a large volume order.
But what we do know is that a very large customer has already spent meaningful time and resources evaluating the technology.
Given that volume production is expected to begin in 2027, I would not be surprised if we get much more clarity over the next quarter or two.
Another interesting development is retrofit.
Historically APR has focused much more heavily on getting designed into new systems early in the development process.
They are now actively looking at retrofit opportunities as well.
Definitely something to watch.
The team is also expanding across operations, quality and sales, exactly the functions you would expect them to strengthen ahead of a manufacturing ramp.
And importantly, management has now confirmed gross margins above 50%.
That matters enormously for the long term valuation framework.
So the thesis remains very asymmetric.
The technology looks increasingly credible.
The commercial pipeline appears real.
Manufacturing capacity is being planned beyond the first factory.
Margins look excellent.
Now they need to prove that qualification turns into real volume.
I’ll publish more of the research later on my Substack, and we will likely also see deeper work from AAIG on their platform.
I’ll also update my original deep dive with the newest information.
For now I’ll link the older piece below.
And if you are interested in liquid cooling more broadly, definitely check out AAIG’s work on the other names in the space. $INV $APRTEC.ST $APRTEC $NBIS $VRT $DELL
Wavester retweeted
Ny presentation av $APRTEC man blir minst sagt lite taggad! 👀🍻🤞
youtu.be/wBahSSKY5sc?is=glt5…
Wavester retweeted
$APRTEC soon testing ATH; The video below is well worth a watch, or a read if you translate the transcript.
Ny presentation av $APRTEC man blir minst sagt lite taggad! 👀🍻🤞
youtu.be/wBahSSKY5sc?is=glt5…
Wavester retweeted
$Aprtec rapport, intressant att de använder Plural när de nämner "det globala techbolag som testar och anpassar Rheion® i sina produkter" 👀 en serie troligen av liknande varianter, olika storlekar? Best case, avtal innan året är slut. 🤞
Wavester retweeted
Presentation från $APRTEC video från robotlinen som tillverkar Rheion, Ai driver ett enormt behov framöver, ett antal NYA kundprojekt har tillkommit. Kassan nyligen påfylld. Italienska Zoppas kommer öppna upp rymdsektorn, tex Leonardo har visat intresse 👀 youtube.com/watch?v=DXEBE3OK…
Wavester retweeted
$APRTEC reports Q2 in five days. Some thoughts going in.
First, we should get a better sense of what is actually being built. The Q2 operational update tightened the structure into three areas, Space & Defence, Energy and Chip Cooling, but there is still a lot to unpack on the product side; the segments got clearer, the products did not.
The relevant information going in, roughly:
Insiders bought. Two board purchases in early July, around 0,4 MSEK and 0,2 MSEK, in the open market just before the quiet period. Small in absolute terms; in a company this size, not nothing.
We are entering the window management pointed at. On the Q1 call the language was clear but hedged; some kind of volume order expected in Q3 or Q4, with an explicit caveat that it could push into 2027. So I see no obligation for anything to land in this report. Best case we already get volume order information here, either from the global tech customer or from one of the smaller accounts. More likely the genuinely critical print is Q3.
Two other things sitting in the same window. The Zoppas framework agreement already slipped from Q2 to Q3. And on 31 July APR completed the DIANA TEVV process and is now prequalified for purchase by all 32 NATO countries via the Rapid Adoption Service; that landed in the middle of Swedish holiday season and I do not think it has been absorbed.
On setup: at roughly 690 MSEK this is still a microcap, and the stock has gone sideways to slightly lower over the past month. The asymmetry looks reasonable. Anything meaningful on volumes and there is real room; a quiet report and I think there is still a bid underneath, because the news flow behind it has not stopped.
Writing actively on the name before and after. For now, my work on who the unnamed tech customer actually is👇
Wavester retweeted
$APRTEC. Let's demystify what the product actually appears to be, because I see a lot of confusion (I was confused too).
Read their website and you get direct chip cooling: 100 to 300 W/cm², hotspot chasing, fluid on silicon. Read their patents in chronological order and you get something else. The recent filings (2023 to 2025) describe one specific object: a factory sealed, non pressurized box holding rows of flat components, a solid state pump circulating dielectric fluid through millimeter channels, a flow distribution plate holding everything within 2 to 3°C, a foil gas bag absorbing expansion so the loop stays sealed for the product's entire life. Zero maintenance, zero moving parts, 20 year design life.
That object is a sealed cooling cassette for BBUs and PSUs. And the industry is about to need exactly that. Racks are going fanless (HPE ships 100% fanless, Rubin trays delete fans entirely), which leaves the power shelf as the last air cooled island in the rack. But you can't just cold plate a hot swappable PSU or BBU module: either you add liquid quick disconnects next to an 800V bus, or you accept lossy dry interfaces. A sealed module solves it: the coolant never crosses the service boundary. And battery modules at high discharge rates need cell to cell uniformity that cold plates physically can't deliver. This is a slot where a pump with no moving parts isn't better than the alternative. It's the only alternative.
The website discrepancy? Companies market the summit and patent the base camp. Chip cooling is the endgame, sealed power modules are the beachhead.
One important caveat: patents publish 18 months after filing. APR filed several families in 2025, during the customer collaboration, that we cannot see yet. It's entirely possible more direct chip level solutions exist behind the curtain. We find out starting late this year.
The PSU/BBU sidecar market is of course smaller than the chip cooling war. But APR's solution fits it unusually well, and the competition is a lot thinner here. Sometimes the best place to start is the slot nobody else can enter.
Wavester retweeted
$APRTEC, serious version.
APR Technologies has completed the TEVV phase of NATO’s DIANA programme and is now pre-qualified for purchase by all 32 member nations through the Rapid Adoption Service.
What actually changed. TEVV is the process through which DIANA assesses whether a company’s technology has reached sufficient maturity for defence procurement. This is not a customer paying for a pilot and calling it validation. It is an independent alliance-level body testing the hardware and signing off that it clears the bar. For a company whose central open question is qualification and manufacturing readiness rather than physics, that signature is the more interesting part of the release.
Why RAS matters more than the headline suggests. The binding constraint for a small Nordic hardware company selling into defence has never been interest. It is the procurement cycle. RAS is built specifically to break that. Allied nations can enter research, prototype or production contracts with innovators without running lengthy competitive processes, and once a prototype has been successfully demonstrated, procurement can proceed without further competition. DIANA can also award contracts on behalf of an Ally through an opt-in programme, reducing administrative barriers. The whole framework sits under NATO’s Rapid Adoption Action Plan, which targets 24 months from identified need to acquisition and integration.
Translation: the path from a single interested defence customer to a production contract is now materially shorter, and it does not require APR to win an open tender against primes with 200 person bid teams.
Context on selection. APR was picked into DIANA’s Energy & Power track in 2024 as one of 73 companies across the alliance. This has been running quietly for two years. It is now finished, and it finished successfully.
What this is not. There is no order, no named nation, no disclosed value and no timeline. Pre-qualification is access, not revenue. Anyone modelling defence contribution today is modelling nothing.
The read-through I care about. Defence is not the thesis. Data center thermal management is. But the reason to hold this name has always been execution risk on qualification, and a NATO validated maturity assessment is direct evidence on exactly that question. It also hardens the second leg: a defence and space channel that is less cyclical, less price sensitive, and structurally more tolerant of high cost per watt than hyperscale.
What to watch next. The first opt-in programme from a specific ally, because that is the event that converts access into a contract. And the Q2 report on 19 August.
Wavester retweeted
Fina nyheter från $APRTEC idag -Förkvalificerad för NATO-upphandling via Rapid Adoption Service- ger samtliga 32 Nato-länder tillgång till en förenklad upphandlingsprocess 👀
aprtec.com/apr-technologies-…
Wavester retweeted
$APRTEC removes the layers.
Their tech, Rheion, runs a special safe liquid DIRECTLY on the chip. No metal plate in between.
And here's the magic: the liquid is moved by electricity, not a pump. Zero moving parts. Uses less power than a lightbulb. xitter.cf/emilhartela/status/207…
Wavester retweeted
Today chips are cooled like this: water pumped through a metal plate sitting on top of the chip.
Problem: the heat has to fight through several layers to reach the water. And pumping harder costs more energy and wears everything out.
It's a brute force solution hitting its limit.
Wavester retweeted
AI chips have one enemy: heat.
The hotter chips get, the slower they run. Cooling is now the wall the entire AI buildout is hitting.
A tiny Swedish company might have the smartest fix. $APRTEC, explained so simply anyone gets it.
Wavester retweeted
$APRTEC på Farnborough Airshow med sin nya partner Zoppas! Tar det som 100% bevis att det är klart med avtalet. PM snart?
linkedin.com/posts/fia2026-s…
Wavester retweeted
Apparently the $SIVE of liquid cooling for AI data centers is also hiding in Sweden.
Long $APRTEC at ~€50M market cap.
My read: this is the only listed pure play on where AI chip cooling is physically heading. The market prices it as a space curiosity.
The bottleneck first. AI compute is now thermally gated.
Rubin class packages pull 2 300W+. GB200 cold plates already run above 50 W/cm². Nvidia's fix for Vera Rubin is brute force: fanless trays, nearly double the coolant flow at the same CDU pressure.
For Rubin Ultra, supply chain reporting points to coolant running directly across the die. Microsoft is etching channels into the silicon itself.
The direction is unambiguous. Coolant is migrating toward the silicon. Room, to rack, to plate, to die, and eventually inside the package.
Now write down what the endpoint requires:
A fluid that can touch live silicon. Water disqualified.
Near zero loop pressure. Bonded stacks take no stress.
A pump at chip scale. Mechanical pumps do not miniaturize.
Millisecond hotspot control. Passive channels cannot.
And boiling inside a package is an unsolved vapor nightmare, so two phase stops at the cold plate.
$APRTEC owns the only commercialized version of that spec.
Rheion: electrohydrodynamic pumping. Dielectric fluid moved by electric fields through microchannels at the die. Zero moving parts. Sub 1W pump power. 74 granted patents plus 31 pending, covering the pump, the microchannels AND the fluid.
Fifteen years co-developing with NASA, ESA and Airbus. ISS hardware delivered. NATO DIANA selected. Claimed 100 till 300 W/cm² and 5 till 10x heat transfer coefficient. Company claims, no third party data yet; the real validation is a paying customer.
That customer: a global tech company paid ~15 MSEK for a year of development designing APR's pumps into its own end products. Pre-series units on volume tooling delivered. Live tests 6+ months, positive per the company.
Disclosed inquiry if tests convert: 150 000 till 500 000 units per year. Hundreds of MSEK annually, from one customer. A separate ~100 000 unit inquiry from data center manufacturers behind it.
Some speculate $SPCX is the unnamed customer. Unconfirmed, but the profile fits: qualifies hardware like a space program, building serious compute.
Meanwhile the sector is being bought empty. Schneider paid $850M for 75% of Motivair. Flex took JetCool for ~$53M. Ecolab just paid $4,75B for CoolIT at 29x EBITDA to align with Nvidia's roadmap.
Two phase remains interesting; $INV via Accelsius might be promising but the holdco structure is less than ideal. And as it happens, the underdog of liquid cooling is also Swedish, just like $SIVE. Sweden keeps producing these strange deep tech chokepoints.
Second act the market has not touched: photonics. CPO is coming and laser frequency is exquisitely temperature sensitive. Rheion is not just a cooler; it is a bidirectional thermal controller that holds a setpoint in milliseconds. Boiling cannot do that. Cold plates cannot do that. Relevant for anyone watching $LITE $XFAB $SIVE.
The board was rebuilt for the ramp in December. CEO from SHT Smart High Tech, sold graphene TIM into the AMD ecosystem, knows this exact qualification grind. NOTE's ex COO on the board, actually ran volume electronics factories. Largest owner Johan Claesson at ~26% with 90+ MSEK invested. Founders ~35%. All locked up.
Production goes 10 000 units this year toward 1 000 000 per year during 2028. Past their own factory ceiling sits a royalty channel (ZIHET, Chillwind). Component economics, not metal bending: ~50% gross at volume, software margins on royalties.
Risks, because they are real:
Single customer concentration.
Venture stage financials (12,5 MSEK nine month revenue).
Dilution near certain to fund capacity.
Claims awaiting public validation.
Undisclosed fluid chemistry, open PFAS question in the EU.
And the stock is VERY illiquid.
I genuinely don't get the price. JetCool cleared $53M as a tuck-in with a shallower moat. $APRTEC trades at roughly that with a deeper patent estate, its own fluid, flight heritage and a paying global customer mid validation.
My guess: undiscovered because it is a Swedish microcap on Spotlight that listed in January. That shifts if the volume order lands.
Position is my own and small. The name is very illiquid. DYOR, NFA.
TLDR: AI compute is thermally gated and coolant is visibly migrating to the silicon. The endpoint physically requires solid state pumping of dielectric fluid at chip scale. $APRTEC holds the only commercialized version, 15 years of NASA/ESA heritage, 74 patents including the fluid, a paying global tech customer in live validation, at ~€50M while the sector consolidates at billions. Long.
Wavester retweeted
"APR – kanske en ny kursraket " Verkar inte vara en betalartikel men de har haft kontakt med bolaget för faktakoll $APRTEC all reklam är bra reklam!
marketz.se/2026/07/12/apr-ka…
