If a Thunderbolt 5/USB4v2 expansion bay module is released for Laptop 16, it would fulfill all the requirements, I think.
And I agree that would be a killer config for many use cases - power efficient on the go while capable when attached to an eGPU.
Theoretically PCIe 4.0 x4 has enough bandwidth for Thunderbolt 5/USB4v2, so we might get lucky and see an M.2 host cobtroller being released, which we might be able to adapt via a dual M.2 expansion bay board.
Problem is, it seems the first “standalone” controller for USB4v2 is going to be released in 18 months best case. This means we won’t see deviced based on it for another 2-5 years after that, and there won’t be any niche M.2 to Thunderbolt host controller boards that’s for sure (as you can’t really find any for TB3 right now).
So much easier and quicker is to wait until AMD supports USB4v2 inside their CPUs…
Didn’t minisforum use one on their strix-halo mini pc? Kneecapped to 64gbit because of the pcie4x4 uplink but till sounds like they had some sort of controller.
Interesting, yeah. Do you have a link to their spec sheet where you saw how it’s connected? Because I’m now also curious how many lanes the “PCIe x16” has there electrically.
No spec sheet but where else are you getting a 64gbit bottleneck from? It’s also quite likely any standalone tb5 chipset will have a 4x uplink just like all tb3/4 and usb4 chipsets had so you’ll need peic5 to get full bandwidth.
Waiting for integrated 80gbit usb4 is probably the best move there.
If framework releases an new amd 13 board with zen6 lpcamm/socamm and 80gbit usb4 (late 2026 or later) I am probably going to be forced to upgrade XD
Almost certainly not 16, probably 4 but with a physical 16x slot. There just aren’t enough lanes to go around.