Pretty sure the rMB will pick up a TB3 port whenever it gets native support in the m line of chips (no idea when that's supposed to be; trying to keep up with Intel chip roadmaps seems like an exercise is trying to make you go insane.)
Intel isn't blocking Apple. It is primarily Apple is blocking Apple here.
"native support" in the CPU package won't help the MB. The CPU package is too far from the port edge for Thunderbolt. The MB suffers from the mismatch between Apple trying to pull all the electronics to the middle of the system, Thunderbolt placement restrictions, battery volume constraints , and being lighter, thinner than MBA.
Intel's TBv3 controllers are all about the same size. 10.7 x 10.7 mm whether single or dual.
https://ark.intel.com/products/series/87742/Thunderbolt-3-Controllers
What Apple needs for the MacBook (if won't adjust the design) is something significantly smaller for single. Perhaps if there was a way to dump the embedded USB 3.1 gen 2 logic and use that from the CPU+PCH package they could shed more size. It is an extremely small space.
https://www.ifixit.com/Teardown/Retina+MacBook+2016+Teardown/62149 ( step 5)
the top ribbon cable here is the 2016 design and the "red highlight" from the iFixIt tear down is primarily just the minimal USB-C switch required for power/usb-c/DisplayPort. You can't get a larger USB subassembly closer because the keyboard (pictured in black below and corresponding battery package in bottom of shell) are jammed almost all the way up to the port along with the mounting screw points for the lid assembly.
( same step 5)
In short, there is too much physical stuff in the corner where the port is. Where the screw mount is or the upper corner of the battery is where you'd want to place a TBv3 controller. If you put the TBv3 controller in the CPU package you have even
less of a viable solution.
"Native" TBv3 in the standard package is very likely to be very similar to "native" Ethernet in the PCH chipsets ( and CPU packages where the PCH is dropped into the package). There is a set of PCI-e lanes that can be flipped into a custom connection to a smaller, cheaper PHY chip that is much closer to the socket. There is still a PHY chip out by the socket; you'll just cut the cost and/or size a bit. The physical constraints for the PHYS to be placed within 1-2 inches of the port will likely remain the same.
When Intel's USB 3 gen 2 solution gets mature and becomes extremely stable in the standard chipset, dropping the USB 3 gen 2 from some of the TBv3 controllers would make sense for ultra small, ultra low power solution targets.
The bigger help the m class could do for the MacBook is drop the overall power requirements so could shrink a corner off that battery so could put a TBv3 controller in there. ( step 3 in the teardown has a larger daughterboard for the headphones and mic assembly on the right, but that runs much cooler than TBv3. )
Or Apple could unpaint themselves out of this corner ( add some thickness and some weight since more metal weighs more weight and therefore have room for TBv3 ).
The MBA is stuck in time and the MacBook is stuck on initial lightness competition with now nonexistant MBA 11". Apple is stuck in a catch-22 of their own inaction at this point. The 2008 MBA only took 2-3 years to kill off the MacBook. Going into 3 years now and MB and MBA still 'resolved' some sort of consolidation yet.
Apple needs to either merge them or move them both forward. One of the two. If keep two then the MacBook could move to 2.2 or 2.3 lbs and still be under the MBA's 2.9.
