After falling into the tinkering rabbit hole with my Raspberry Pi Zero ages ago, I’ve built dozens of useful projects. And with the Raspberry Pi 5’s release a few years ago, there was no turning back. Since the RPi 5’s hardware surpasses its predecessors on all fronts besides hardware transcoding (that’s a tale for another time), I could experiment with tons of cool DIY ideas without running into performance issues.

Unfortunately, as my projects got more advanced, I started to notice that the Raspberry Pi had trouble keeping up with my experimentation needs. Sure, it could double as a Kubernetes node, a FydeOS-based Chromebox, and a retro gaming station. But compared to an x86 mini-PC, the Raspberry Pi’s lack of computing prowess became quite evident – to the point where I’d rather create certain projects on a cheap mini-PC than on a Raspberry Pi.

General-purpose Windows 11 system

Botspot VM is the closest you'll get on the Raspberry Pi

Despite the Raspberry Pi 5’s compatibility with essential Linux distributions, many operating systems remain out of its reach, including Microsoft’s flagship OS. As much as I like throwing flak at Windows 11, I was tempted to see how well it would run on the RPi 5. After all, certain tools, like Adobe’s suite, PowerShell, and Visual Studio (not VS Code; that’s entirely different), tend to remain out of my reach on Linux, and the last two are especially important for my scripting needs.

But since Windows 11 isn’t officially available on the Raspberry Pi, I had to go through different methods that used a bunch of workarounds. A bare-metal installation is nowhere near good enough, since it’s a buggy mess and is full of compatibility errors. A Runtipi-based Windows 11 container works, but it’s still not good enough. Botspot's virtual machine is a surprisingly fun method, but considering it’s only a VM, I can’t really use it for advanced scripting projects.

Meanwhile, most of the mini-PCs I’ve tinkered with are more than capable of running Windows 11. Even the lightweight champ N100 can run most applications, including nested VMs and WSL setups, fairly well. Now, the Raspberry Pi is good enough for learning Linux – heck, that’s how I got my distro-hopping obsession to begin with. But if I’m looking for a small device that can help me practice Adobe’s industry-grade Creative Cloud, hardcore PowerShell scripting, or extensive .NET coding, I’ll always go for a mini-PC instead of a Raspberry Pi.

Home Assistant hub

Especially when you need several add-ons, integrations, and automation chains

Let me preface this section by adding that my 8GB Raspberry Pi 5 served my smart home well as a Home Assistant hub for several weeks. But once my collection of smart devices and their automations started to go up, HASS got noticeably slower. The situation only got worse as I started tinkering with community-created blueprints, HACS integrations, and (most importantly) add-ons. Slotting an SSD helped improve the performance and made my HASS setup more resistant to constant log generation, but I could no longer accommodate more IoT paraphernalia into my smart home without lowering the performance of the RPi.

An x86 mini-PC, which costs slightly more than the high-end Raspberry Pi 5 + SSD combo, easily surpasses the SBC when it comes to HASS workloads. With a cheap mini-PC, I could run dozens of automations involving different smart devices – and I don’t mean HASS’ trigger-action chains, either. With self-hosted add-ons like ESPHome and Mosquitto Broker, I could bring different IoT gadgets into my control hub and automate them alongside entities imported via HACS integrations using a local Node-RED instance – all without turning the Home Assistant UI into a slogfest.

Secondary Proxmox node

For a high-availability PVE cluster

While I tinker with different server platforms, Proxmox is the one I prefer more than anything else. Since Proxmox supports high-availability clusters, deploying one is a sure-shot way to ensure my essential services remain operational even when my primary workstation goes down because of my experiments. Even leaving aside the fact that I can’t expect x86 LXC and VM images to work on an Arm-based Raspberry Pi, I can’t even use the SBC as a standalone Proxmox node, as PVE doesn’t have an Arm image.

And no, my attempt at running the unsupported Arm variant of Proxmox doesn’t count, either. Between a plethora of compatibility issues and the low CPU firepower of the RPi, it’s clear that the SBC isn’t good enough for normal PVE loads. In contrast, a mini-PC can double as an amazing LXC-hosting workstation, especially since Proxmox has fairly low system requirements. Heck, the secondary nodes for my current high-availability Proxmox cluster comprise an N100 NAS-cum-mini-PC and an N150 x86 SBC, and they can reliably keep my important self-hosted stack running at all times.

Self-hosted router

You’ll want at least two network ports on the mini-PC, though

Over the years, I’ve moved from one custom router OS to another, with a Raspberry Pi-based OpenWRT firewall being the first of the many tools I’ve experimented with. Although my Raspberry Pi served as a great learning resource for my cybersecurity escapades, it proved inadequate for advanced routing, VPN encryption, NAT traversal, and other demanding operations. Then there’s the fact that I had to use an external USB-to-Ethernet adapter as the WAN interface, which is far from ideal.

Switching to an x86 system was a great investment, as I could go for something as hardcore as OPNsense. OpenWRT is the only router OS that works on the Raspberry Pi, unless I go through some workarounds that, as you’ve already seen, would be more trouble than they’re worth. Not only that, mini-PCs are better suited than Arm-based Raspberry Pi boards for the extra services I tend to enable on OPNsense. Plus, many of the sub-$200 (or even $150, in some cases) mini-PCs tend to ship with dual NICs, so you don’t have to deal with adapter shenanigans like I did, either.

Nevertheless, the Raspberry Pi has its own use-cases

So far, I’ve only gone over the projects where mini-PCs curb-stomp the Raspberry Pi family. But there are certain DIY ideas that pair better with these Arm SBCs. Projects involving GPIO pins, for example, work way better with Raspberry Pi boards, and this includes everything from circuit components to sensors, IoT peripherals, and HATs. Likewise, if you’re planning to run lightweight containers without expending too much power, the newer Raspberry Pi boards might be right up your alley. Source? Yours truly, who has an entire monitoring stack running on a Raspberry Pi.

👁 Monitoring a CasaOS container in Uptime Kuma
I self-host Uptime Kuma on my Raspberry Pi to monitor my home lab - here's how

Lightweight and easy to configure, Uptime Kuma lets you keep a weather eye on your home server