Generally speaking, you won’t find forward-looking statements about what to expect in the future regarding software (or hardware) products by NVIDIA in these forums. These forums are primarily intended for community based support of current products, in their current state. These forums are not generally where NVIDIA publishes information about future products.
Regarding CUDA releases historically, if you study what NVIDIA has done in the past, you will find that in nearly all cases, updated or new support for new CUDA development platforms (or newer toolchains within existing platforms) comes about via a new CUDA release (either a major release, like 9.0, or a point release, like 9.1) Based on that historical pattern, a likely place for NVIDIA to introduce new support would be in some future CUDA release, whenever that may be. I’m not able to disclose publicly on these forums when that may be; refer to the first paragraph above. I’m also not claiming that the next release will contain any changes in this respect. But if NVIDIA were to offer additional platform support, it would likely come via some future CUDA release.
If you are a representative of a company with an already-executed NDA with NVIDIA, then you may be able to get more information through that relationship. However that information won’t flow through these forums, whether publicly or via private message. So there is no point sending me a private message about it. I likely will not be able to respond meaningfully to such messages.
Currently, many folks have been able to use CUDA 9.1 with the 15.4 toolchain in MSVS 2017. You can find a number of related messages and threads on this forum concerning that.