This one.
On a side note, there appears to be some speculation that Google is using 3 different versions of G3 chips in the pro, standard 8 and 8a. I assume different bins of the G3 chip.
Copied the text from the post...
Seems like google split tensor G3 into 3 versions by different packaging:
G3 Lite iPoP { P 8a }Interposer Package on Package
G3 Fo-PLP { P 8 }Fan Out Panel Level Packaging
G3 Pro Fo-WLP { P8 Pro }Fan Out Wafer Level Packaging
Last edited: