Is 8GB of RAM is Enough to Run Qubes Comfortably?

Hey, I am getting a thinkpad specifically to run qubes. I found one, at a great price (I have a tight budget), with 8gb of ram, problem being that this specific model has the ram soldered in, so I can not upgrade it in the future. With that in mind, I was wondering if 8 is enough to run qubes comfortably if I am not planning on running dozens of vms simultaneously. Specifically, I am planning on running the following vms at the same time: sys-net, sys-firewall, sys-vpn, sys-whonix, anon whonix(general browsing), vault vm(keepass) and a fedora dispvm (browser, no streaming). Is 8gb enough or will it be too slow? I believe it’s enough for them to run but not sure if it will be too little to use the laptop comfortably.

If I recall correctly I used to run Qubes on a laptop with 8 GB RAM with a similar workload. There weren’t any problems except I wanted to run more VMs simultaneously so I upgraded. If you think you’ll be happy running that amount of VMs simultaneously until you buy a new computer, then I think you’re going in the right direction. I’ve considered soldering my RAM onto the motherboard to defend against ‘cold boot attacks’ but it’d prevent me from reusing the RAM for a future project and cold boot attacks aren’t much of a threat to me. For overall performance there are other things to consider such as storage device speed that can affect the time to open and close VMs, boot and shutdown of the system. I think it’d be a good idea to take a look at the HCL to confirm that the laptop is compatible with Qubes before buying it.

I don’t know how small your budget is but I would not recommend buying
a model where RAM cant be upgraded. If you have even a small amount free
in the future you can upgrade and get a better experience. If you go for
the soldered in option, that isn’t open to you.

8GB would do (just about), but you would have to be careful about
allocation. This is a topic that has come up before here and on the
mailing lists.
I suspect that Whonix will grab a lot of that RAM - there are
alternatives if you find it too slow.

(changed the title to be about the RAM limitation as that’s the key issue of the thread)