That seems like a really tight latency budget to run through even an optimized wireless stack.

Working backward: frame needs to be received by headset in time to refresh, which means it needs to be sent by host device in time to make its way through the headset wireless phy and be decoded.

And working forward: headset needs to encode sensor values, send across wireless, be decoded on host, before frame can start to be generated.

Which leaves the time in-between those for actually doing all the processing and frame generation.

Ditching the wireless phy and low-level steps seems like an easy trade-off for more processing time.

Even wired that seems like it’d need a lot of copper. The power cable connector right now is pretty pliable. Most thunderbolts cables are not as much. And who knows if it would need to be even thicker than that.

You could do something like an active optical cable if you wanted - adds expense but given the price tag of this, probably not insurmountable.

Well, I do now. I'm hungry.