I might be missing something, but unless you're talking about the race-prepared, straight-cut teeth transmissions, the DCT is faster than the BMW SMG's that were used on the E36 and E46 M3 (single clutch automated transmissions) in shifting. They both have two clutches and helical gears.
The BMW SMG I/BMWSMG II/Audi R Tronic/Lamborghini e gear etc. transmissions all have shift times in the 150-300ms range and are basically dead because they're painfully jerky and slower than dual clutches, although they perhaps enjoyed a heyday in the late 1990s-2007 period before DSG/DCT came around. The Lamborghini Aventador is probably the last production car to have one, after the smart fortwo finally got a dual clutch.
Although jerky shifts feel more "real", it doesn't necessarily mean faster performance. I drive three pedal manual but would not put the ZF 8 into the "soulless mass produced commuter transmission", simply because the shift speeds are so good, even if it is super smooth. Then again, I contrast with ancient 3/4 speed autos and how 1-second long downshift reaction times would be sold as "tiptronic". I agree with chassis on the "in house not always better" - mainly based on Mercedes experiences as well, although with the 2-year-older 7G-Tronic, which was significantly slower and jerkier than I hoped for such a modern car. Also, Porsche somehow managed to acquire a 5G-Tronic on the 911 for a while, so BMW could license out one of the exotic transmissions if they really wanted.
You can watch videos of the smart fortwo 0-60 single-clutch automated manual, it's comical how long the shifts are.
If there's one single transmission that killed the stick shift, it is indeed the ZF8. You now need to sacrifice 2-3MPG's and 0.2s usually in the 0-60.
One anomaly I've found is, if the 7DCTs are so good, why do they still get 1 MPG worse than the 6MT on the EPA tests? M2 DCT is worse MPG, M4 DCT is worse MPG, so weird.
Happily for shareholders and unfortunately for sports car enthusiasts, companies all exist to eventually produce profits and results for their shareholders. When the halo effect and branding power of subsidizing your low-volume sports cars with your high-margin SUVs pays off in higher SUV sales, then you do it. When siphoning off some of those profits to R&D a slightly more technologically advanced cargo box in the X5 trunk or better air suspension or 360-degree cameras for the large vehicles is more worth it, executives will direct the money there.
BMW only makes about ~$5k profit per vehicle, Porsche makes $15K and has all of VW AG to share tech with, so there is less wiggle for BMW. Even so, the vast majority of Porsche profit is purely due to the badge and selling boatloads of Cayennes and Macans.
I'm not sure what the exact R&D cost of a new DCT would be, but the volume of DCT usage is surprisingly quite low for BMW. Making it work everywhere is not even that worth it to Audi with all their parts sharing sometimes, as they stick the S5 and RS5 with the ZF 8 as well since it accelerates just as nicely and they didn't have any R Tronic/DCT that fit the torque requirements, and works in the monstrously powerful Hellcats too. I wonder if there's a fundamental mechanical trade off in that a torque converter is a force multiplier while a clutch is fundamentally a torque divider, that gives torque-converter autos an advantage on takeoff or in shifts.
I'm also still a little surprised that it's so hard to just program a DCT to be ultra-smooth (can't you just rev match perfectly? There is already some DBW delay anyway), but what do you know.