It would require an entire separate TPM chip, integration of it on the main PCB, and all the the firmware and software handling that comes with that, and collaboration with the GrapheneOS team (which I hear on forums and people who have worked with them, is often not a pleasant experience) for an extremely small percentage of their sales.
Doing /e/ or calyx would definitely be significantly easier.
It would require an entire separate TPM chip, integration of it on the main PCB, and all the the firmware and software handling that comes with that, and collaboration with the GrapheneOS team (which I hear on forums and people who have worked with them, is often not a pleasant experience) for an extremely small percentage of their sales.
Doing /e/ or calyx would definitely be significantly easier.
I hear all of that, but I’m going to maintain some hope since otherwise I’m stuck on google hardware for the foreseeable future.