Given that valve refuses to use KLA for their own competitive multiplayer games, and has gone out of their way to not make their hardware locked down, I really dont think they will go down the path of making a locked down platform or facilitating intrusive anti cheat.
A TPM is entirely under your control. It's designed in such a way that you can't do certain things with data within it, but that's not because (at least in theory) someone else can and is controlling your TPM to prevent you from doing those things. The TPM, unlike an installation of Windows, doesn't only listen to Microsoft.
Similarly, the presence of unsigned software on a computer would not stop a Linux kernel level anti-cheat from working, and the kernel level anti-cheat shouldn't prevent the unsigned software from working. Once you run that unsigned software, your machine is tainted similarly to the way your kernel is tainted if you load the NVidia driver.
There are several products that rely on a USB device like this for DRM solutions. It’s probably much easier to unlock static assets than validate running code, but I don’t have insight on the true complexity.
What does the USB stick actually do? The hard part of implementing the anti-cheat (ie. either invasive scanning or attestation with hardware root of trust) is entirely unaddressed, so your description is as helpful as "would it be possible to implement a quantum computer as a USB stick?"
They have bet on the behavioral analysis anti-cheat horse but it hasn't won any races yet.
Moreover, they've proven that it's certainty more difficult to get it working than regular old fuck-the-end-user anti cheat.
Lastly, don't assume that the success of the platform will persuade these companies. They were already firmly un-persuaded when the steam deck got popular. And really, I think the popularity of a platform depends on the support of these companies more than the support of these companies depends on the popularity of a platform.
Depends on just how successful SteamOS gets. If it start to have a significant market share, competitive multiplayer games might start to find it hard to ignore it. Though how they decide to deal with that, I have no idea.
I think Valve see a future for anti-cheat where most of it is behavioral analysis. Client-side anti-cheat is a big game of cat and mouse. It does make cheat harder to develop, but to a point where the customer is impacted. Post game analysis cannot be countered "technically". Cheat would need to mimic a real player behavior, which at the end is a success. If you can't tell if a player is cheating or not, does it matter that they are ? Although for things like wallhacks, it might be harder to detect.
This is basically effectively where KLA has gotten to. There are still plenty of cheaters, people just don't realize.
I think it does matter in a strictly moral sense, and if people were more aware of how bad the problem is, they would likely be outraged. Alas, since they can't see it, they are not aware of it, so there is no outrage and the games companies are satisfied.
They use different means to detect cheaters, which means sometimes they are banned several weeks after the fact, but they do ban cheaters.
* I don't care what the intention is, they are _objectively_ intrusive.
** Last time I argued this, someone seemed to assume that I was claiming that writing Linux kernel drivers is harder than Windows kernel drivers. I am not arguing that, you need some kind of trusted party enforcing signed kernel drivers and a signed kernel in order to make KLA sufficiently hard to bypass.
*** In terms of the average Joe just wanting their game to run rather than having to think about the ethical implications of buying hardware you don't actually own or running an OS which gives control of your hardware to various corporations (but not you).