IMO, requiring a TPM for any kind of attestation wouldn't do much because they can be procured in the tens of thousands for not much money at all. Then they use an SPI bus to communicate, so you could basically build a cheap device that only multiplexes dozens, hundreds, or thousands of TPM on a single physical host.

The real sham of this, to me, is that Google's talking nonsense about ensuring the client device is "trustworthy" for whatever their criteria means. But in reality the client needs a real assurance that the site it's visiting isn't malicious, serving malicious content, or otherwise collecting data that could be used for malicious purposes. Google has directly failed two of those three for many years, and one of them is their entire business model. Where is our protection from Google?

Maybe Google should use their clout to work against DRM online, and push back on the insatiable corporate greed of most of the content creation corporations? Especially those busy cutting down trees to prevent striking workers from getting shade?

Adding on to this, what of people in sanctioned nations? Google, as a US entity, is compelled to adhere to US law and to sanction nations that the US deems should be sanctioned. What about activists in those nations? What about targeted populations in those countries? What happens when a minority group is targeted by a hostile government and that government demands logs of device tokens accessing information the government doesn't like? This idea is nonsense on so many levels, and such a 180 degree turn from how the internet has developed over its existence.