"Uncompromisingly secure", yet contains an Intel CPU with an unpatchable mystery blob. Just put an unlocked ARM CPU in there already.
There are good ARM CPUs, like in the NVIDIA Orin modules, or in some boards offered by MediaTek and Qualcomm, which are derived from their smartphone products, but all these are very expensive in comparison with the Intel products.
If you are not scared by much higher than Intel prices, you can look e.g. for products including these:
https://www.mediatek.com/iot/products/genio1200-devkits
https://www.nvidia.com/en-us/autonomous-machines/embedded-sy...
(at the NVIDIA site you must ignore any module or devkit that does not include the word "Orin" in the name, everything else is too old junk, even if it appears to be cheaper)
For cheaper than Intel, but with performance at the level of the previous Intel Atom CPUs (i.e. Tremont-based, like Jasper Lake and Elkhart Lake) there are many products using the Chinese Rockchip CPU RK3588, for which much more public information is provided than for any Intel or AMD CPU, so I would trust it much more for implementing a secure appliance like a firewall, but not all of those products are well designed and some are overpriced (i.e. anything more than $200 for a complete computer with RK3588 is grossly overpriced).
The advantage of ARM in laptops are clear - the battery life/performance/heat tradeoff is better. It’s also clear for server farms.
But if I’m just running Linux on a desktop anyway, why would I care about ARM vs x86 unless I’m targeting my builds for other ARM computers.
The "uncompromised security" line suggests that the product is targeted at this group. The Intel CPU shows that it's actually aimed at people "just running Linux on a desktop anyway".
On the other hand, for a very cheap, secure and fanless (i.e. reliable and silent) network appliance, like a firewall/router/NAT/Web server/e-mail server etc., a small ARM-based computer, e.g. one with RK3588, may be the best choice. (but not all ARM CPUs are documented well enough to make them more trustworthy that the Intel/AMD CPUs)
If your insecure personal computer is shielded from the public Internet by a secure firewall, you need not worry that its remote management features, which are supposed to be disabled, can in fact be activated without your knowledge.
And that would depend on the processor more than anything else. E.g. raspberry pis consume more power on idle than some of these x86 boxes which are more powerful (but more expensive)
As if that existed.
Also, I ponder what is the threat vector presented by an "unpatchable mystery blob" that is not already present without one.
Let me explain what unique threat the blob poses:
https://securityaffairs.co/58656/hacking/intel-management-en...
I thought you were talking about proprietary microcode, which all of these chips you mention do come with. Or even worse, since they are all SoCs.... You'd have to e.g. avoid using the Ethernet controller on them (which generally has flashable eeproms with code) which would also happen to be an effective prevention of any ME issues even if were running an ancient version of ME.
I.e. the only threat vector you mentioned does not apply and can't possibly apply. A mac shares more firmware with your average intel laptop than this.