Louwrentius

I built a 10-inch mini rack from aluminium extrusions and I had a lot of fun doing it. I want to share my build in this post.

In January of 2025, Jeff Geerling released a video about 10-inch mini racks. I was absolutely oblivious to this new trend, and I instantly knew that I wanted to build one for myself some day, although I didn't have a real use for one.

That said, I've been working on a virtualization project recently and I've bought six 1L PC's (three for each simulated datacenter).

Because these 1L PCs are small, they have huge external power bricks. As these mini PCs are also connected to two networks, they create a huge mess on my desk. Finally a reason to build a mini rack, to tidy things up!

There are a few different brands of mini rack for sale, but I wanted to make one for myself. Prebuilt racks (kits) are not cheap for what they are so I wanted to try and see if I could build one myself for less money. Turns out you can buy 20mm aluminium extrusions and accompanying components to build your own rack.

(Aluminium extrusions have a standardized 'groove')

Aluminium extrusions are bars with a groove on all four sides. These bars have a standard format and you can slide all kinds of equipment in there and lock it in place with set screws. It seems to be used a lot for home made 3D printers, CNC machines and whatnot. In the picture above a special corner piece is used to connect three bars together, fixed in place with set screws.

The L-brackets can be used to create T-sections within a frame to sturdy the structure and provide additional mounting points. In my rack, the middle post carries the back of the shelves holding the mini PCs.

These sliding cage nuts (M5) can be used to attach anything anywhere. In this example we used four of them to hold the side panels in place. These cage nuts can also be used for their intended purpose: mount 10 inch rack-mount equipment.

In this picture below, some cage nuts are left that hold the side panel in place. Also notice in the upper left that I've used cage nuts to attach some black cable tie holders that in turn keep cables in place.

The computer trays

I planned on ordering 10 inch 1U shelves and be done with it. Unfortunately these metal shelves are too expensive for my taste and would have cost more than the aluminium frame including components (I need 8 shelves).

Many 10-inch rack builds - such as the ones featured by Jeff Geerling - use 3D-printed face plates to mount various kinds of equipment. Jeff showcased some of these models in separate videos. That said, I decided against using 3D printed shelves.

First of all, I don't have a 3D printer and as useful as 3D printing can be, I feel that 3D printers often turn plastic into landfill. I'd probably feel differently if (more) sustainable materials would have been used¹.

So instead, I chose to order cut-to-size aluminium sheets and I used the L-brackets to hold them in place.

The aluminium shelves turned out OK, but they are not ideal. The 1mm thick plates do bend slightly under the weight of the computers, although it's still fine. Aligning the four L-brackets on the same horizontal plane was a pain. Filing off the sharp corners of each plate was no fun, I should have ordered them with rounded corners.

I think these aluminium plates create an open design that is better for keeping the machines cool. There are many 10-inch self 3D models available for these 1L PCs but they all create a tight collar around the front bezel of the computer, which looks amazing, but I don't think it's great for airflow. Regular metal shelves would also have been fine.

10 inch power distribution

As you can see below, the backside shows the internals of the case are a bit of a mess². The truth is that I've should made the rack at least 1U higher to accommodate the very lengthy cabling of the power bricks. All the cabling does fit, but it's not easy to make it clean looking and also give the power bricks - which lie at the bottom - some airflow.

No price for cabling management and neatness, that's for sure

I've used two 10-inch rack mount power distribution units from Brenenstuhl. They were cheap but they unfortunately didn't fit in a horizontal position. The PDUs are the only actual 10-inch rack-mount component in the entire build and the fact that they didn't fit felt ironic.

The cause is simple: in a 19-inch or 10-inch rack, the square holes holding the cage nuts are 'flat sheets', so the power cable sticking out of the side of the PDU can flow behind those square holes. If you use the 20mm aluminum extrusions, there is a 20mm bar in the way. This is why I had to mount the PDUs vertically, which did work fine. At the top, the PDUs are kept in place with another L-bracket clamping the PDUs firm against the rail.

The external power bricks of the 1L PCs' are a huge pain. Having six cords and external adapters was absolutely no option for me, I wanted a fully self-contained rack. If you ever build your own rack, try to use computers that have a power supply build-in.

I wish I could power these 1L PCs with a 'power shelf'. In this case, a power supply with enough capacity to sustain these PCs at the required voltage and with the proper brand-correct power jack, maybe like this.

Cooling

All power sockets are in use by the six 1L PCs and the two network switches, so how are we going to power the two cooling fans³?

Fortunately there exist USB-to-fan-header cables. It feels dirty but it isn't. USB is only 5 volt but these adapters contain a boost converter that outputs 12 volt for the fans. The power draw of these fans (around 1W) is well below the threshold of 4.5W for a USB3 port.

The open backside probably doesn't help with cooling and Ideally I'd fill in the gaps. Due the the irregular shapes, I feel it's too much effort. I do feel a proper airflow at the front of the 1L PCs, so I think it's 'certified good enough⁴'.

Networking

The 1L PCs have two network connections, one to each switch. The 3Com switch at the top is 1 Gbit, the TRENDnet switch is 2.5 Gbit. I used the Conceptroning ABBY12G USB3-to-2.5Gbe adapter to connect the 1L PCs to the switch.

The 2.5 Gigabit network is a backend network for live migration of virtual machines amongst other traffic. I'm able to achieve line-speed with both iperf3 and virtual machine migrations.

The gigabit 3com switch is at least 26, maybe 27 years old!

The rack has become a bit crowded due to all this network connectivity, as seen on earlier pictures. It didn't help that I ordered short UTP cables that are quite stiff.

Panels

The side panels are sheets of anodized aluminium which look good, in my opinion. I forgot to order panels for the top and bottom. Therefore, I decided to go with some wood panels instead of aluminium sheet metal.

A local shop offers scrap wood for peanuts and also cuts it to size for a 'few peanuts more'. Very handy if you don't have the right equipment to make straight, clean cuts of wood.

Although not pictured, the bottom panel is kept in place with screws locking into cage nuts. These screws also hold the rubber feet in place.

Problems

A lack of access to the VGA/DP ports:

If I lock myself out of a machine, I need to remove one or more fans and connect a VGA/DP cable to the back of the affected machine. This also means moving a lot of UTP cables to the side. I can attach a keyboard to the front but attaching a monitor is thus a real pain. In more regular rack builds, you can add 'keystones' that extend ports to a panel on the outside of the case, where you can connect to an interface without any issue.

Cost

All prices include 21% Dutch sales tax. The total price excludes the two Noctua fans, which would probably add another 40 euros.

The price is a bit inflated because I overbought small items for rack building. If I opted for wood paneling instead of aluminum panels, that would probably also cut the paneling cost in half.

How I use this mini rack

The mini PCs are running Debian and they all act as virtualization hosts, using KVM. The gigabit ports are used for the management and provisioning network (PXE+TFTP+iPXE+HTTP). The 2.5Gbit network is used for virtual machine migration and a VXLAN network that encapsulates all the different virtual machine networks.

By default, this rack is off. When I want to use it, I turn on a Zigbee power adapter and after a delay, wake-on-lan packets are sent to all six machines to power them on.

Power consumption

The two switches, two fans and six 1L PCs together use around ~90W idle.

Evaluation

I'm quite happy with this build. Cost was acceptable, I think it looks decent enough and it really cleans up my desk. Although cable management is clearly not my strong suit, I feel it's an overall improvement. Maybe a few handles for carrying would be a nice future addition, but for now this rack is finished.

Did I save any money? I'm not so sure.

Hacker News discussion

Acknowledgements

• Jeff Geerling for introducing me to 10 inch racks
• This build by Logan Marchione inspired me to further look into aluminium extrusions

Disclaimer: I'm just a random person on the internet and this article is my personal opinion. This article was written without AI.

The first time I heard from Rutger Bregman was due to his remarks at Davos about taxing the billionaires. His speech went viral and I was rooting for him. Later on he got to troll Tucker Carlson about Carlson being a millionaire working for billionaires. Absolutely fantastic.

A few books in, Bregman started the School for Moral ambition in 2024. Although he is a co-founder, he seems to be face of this non-profit organisation, doing a lot of high-visibility interviews and he's the author of the book "Moral Ambition".

The idea behind the School for Moral Ambition (SMA) is quite simple: use your talent to address the world's biggest problems. To paraphrase: "don't become a consultant at McKinsey but do something more valuable and worthwhile with your talent". The argument is similar to what Steve Jobs said to John Sculley: "Do you want to sell sugar water for the rest of your life, or come with me and change the world?"

It sounds all very noble, but I've become a skeptic for multiple reasons.

Bregman rubbed me the wrong way in the last podcast episode of the "De Rudi & Freddie Show" in 2024. He confronted co-host and award-winning investigative journalist Jesse Frederik and openly questioned if his work really made a difference. Bregman implicated that Frederik could do other things that would be much more effective and be more impactful than his investigative journalism (I'm paraphrasing here).

I'm absolutely comfortable with questions like these, although it can feel confrontational. Yet I really felt it came from a somewhat righteous and smug moral superiority. I felt this way because I haven't seen any evidence¹ that SMA has done anything of substance that warrants all the money involved.

Bregman drank the AI cool-aid

Now I already learned that Bregman was really enamoured with AI, and I'm absolutely not, for the usual reasons. That said, fast forwarding to today, I truly believe he has lost the plot, and this recent video (June 2026) is evidence of it.

He is comparing skepticism of artificial intelligence with climate change denial or climate skepticism. He sees AI as a tool that is beneficiary, just not in the hands of billionaires but in the hands of governments (as I understand it). He believes that AI will create an utopia where people can actually work less hours as Keynes 'wrongfully' predicted

Unfortunately for Bregman, as far as I can tell, there is no (scientific) evidence that AI is actually capable of having such a huge impact. Meanwhile compare this to the mountains of evidence for human-caused climate change, it would be a different story. I would expect a different argument from a former journalist. To reiterate: the case for climate change is rooted in a vast body of scientific research. The case for AI creating a 15-hour work week seems mostly backed by wishfull thinking.

I feel that Bregman isn't really aware of his own biases, fuelled by his own personal experience using AI. He paints all the AI critics as 'luddites'² and that AI-skeptics are on the wrong side of history, just as the climate deniers are. Comparing AI skeptics to climate change skeptics feels like a cheap rhetoric trick that doesn't do justice to the criticism and skepsis surrounding AI.

Bregman also believes that stopping AI or 'shutting it down' doesn't make sense. Because if we shut AI down in our countries, other countries or regimes will build AI and get an advantage over us (he shows China as an example). The age old "if we don't, somebody else will", that has excused bad behaviour for all of eternity.

The reality is that voters and the larger public never had any say how this AI technology is deployed. That we take into account the impact on communities, people, the use of energy and other resources. We never had any say in this. The entire AI-push seems anti-social, anti-democracy.

Yes, the companies now pushing AI upon us are so huge and valued so high they are dwarfing the GDP of entire nations within Europe as Bregman suggests. Yet, I think there is strong evidence these numbers are not real. I sense such a lack of critical thinking and healthy skepticism about his own claims. He is dismissing the work of people like Ed Zitron who has unearthed a ton of evidence that AI doesn't deliver (or not nearly as much as is widely claimed) and there's a bubble that may collapse at some point. As I see it, Bregman only has a well-produced narative, but nothing more: it seems pure rhetoric.

And then I learned something about SMA that made me really disappointed.

The School for Moral Ambition took money from Bill Gates

Bregman has been quite positive about Bill Gates contributions to society, mainly pointing at his work on fighting malaria. Unfortunately, it seems that Bregman has not done his due diligence about Bill Gates and his foundation.

Fortunately, Tim Schwab - an investigative journalist - has investigated the Bill Gates foundation and the result has been documented in the book "The Bill Gates Problem: Reckoning with the Myth of the Good Billionaire". In this book, Schwab has demonstrated how problematic Bill Gates and his foundation really is. Remember that Bill Gates had deep ties to Jeffrey Epstein, which was cited by his ex-wife as a reason to file for divorce.

Tim Schwab had contact with Bregman about this and I think it doesn't look good for Bregman. It's really bad for a former journalist to cite a source as evidence for Bill Gates good work that has financial ties to Gates. Schwab repeatedly tried to contact Bregman about this, but got the statement through Bregman's agent Anne Strunk:

“He’s not a fan of your style of journalism, which seems grounded in bad faith instead of curiosity,” Strunk noted in an email.

When Schwab wrote his article in 2025, it wasn't known that Bregman and his School for Moral Ambition was funded in part by Bill Gates, Schwab discovered this by accident. (It's now disclosed on the website at the bottom).

Now for the life of me, I can't fathom boasting about moral ambition and accepting money from Bill Gates. I just can't. And they know it looks bad because if an organisation omits (at first) mentioning receiving funding from Bill Gates, you know they know how bad that looks. Lying by omission.

If SMA had any actual moral ambition, I think it would have been a cooperative. (Now it's just a typical hierarchical and opaque organisational structure with the 'owners' including Bregman, at the top.) I would expect that this organisation would also be extremely transparent where the money is spent on exactly. The impact report didn't really clarify anything for me. It's all testimonials, stories and anecdotes.

It gets so much worse

My jaw dropped when I read this article by Anita Naidu. I learned that Dutch journalist Marieke Kuypers discovered that the USA branch of SME runs out of the offices of crypto billionaire Mike Novogratz and that Novograt's sister sits on the SMA's board. Mike Novogratz donated a milion dollars to Trump.

At this point, I think that both Bregman and SME are absolutely unredeemable.

Accusations of racism and elitism

In interviews I've noticed that Bregman talks a lot about 'talented' people. For him, talented people are highly educated people. To get a fellowship with SMA, there are all kinds of criteria a candidate needs to hit and one of them is academic credentials, work experience and so on.

A picture does arise of a very elitist, and also tightly controlled organisation where the founders determine - based on criteria they pulled out of thin air - if you are a good candidate or not. And they alone decide what is moral, which topics should be spend time and effort on.

This is what Anita Naidu articulates in an article from 2025 where she states:

Let’s be clear about who it’s built for. SMA isn’t for people organizing at the grassroots, risking safety to confront systems of oppression. It’s for the already credentialed: Ivy League grads, McKinsey consultants, social entrepreneurs. People groomed to lead—just not to surrender power. The School doesn’t ask them to reckon with injustice. It teaches them how to perform reckoning without threatening their position.

She continues later on:

Instead of interrogating capitalism, colonialism, or white supremacy, SMA asks: how can we make these systems feel more ethical? How can we reframe domination as duty? Structural violence becomes a backdrop for personal growth. Justice becomes self-improvement. Revolution becomes résumé fodder.

Meanwhile, Payal Arora wrote a critical piece on Bregmans article (an adapted excerpt from his book 'Humankind' ) written in 2020. She critisizes Bregman for simplifying the story, leaving out important context about colonialism and racism, showing that the story is actually a counter-example to his hypothesis that people are good by nature³.

According to Payal in this article, because of her criticism, she got into a 'twitter war' with Bregman about her criticism.

I also found a critical article and a follow up article from (yet again) Tim Schwab about how Bregman handled the uncovering and critique on hist past racist remarks about Rihanna.

Closing words

At this point I can only ask the reader not to get involved with the School for Moral Ambition as I think SME isn't moral at all. I believe there is nothing of substance to SME and I don't see anything tangible this organisation has achieved with the money spent.

Disclosure: I have no affiliations with tobacco producers or processors, I'm not a wealthy person trying to evade taxation and I'm also not in any way affiliated with food or meat production. In short, SMA isn't working on anything that would affect me personally in a negative way.