I’m not sure, and don’t feel confident in predicting the future. There are multiple possibilities. 
Right now, the old Alexa skill is scheduled to go away in a month or so and the new Alexa skill has two known issues which are not actually bugs:
One) it exposes all smartthings connected devices— It doesn’t allow you to specify which ones, which the old skill did
Two) doing this creates duplicates for any devices which are also connected via their native integration.
Smartthings engineering is aware of this, and seems to consider 2) an issue that should be fixed, but 1) Less important. There was much discussion of this in existing threads so I won’t go into all of it again.
In addition, there are three known bugs that occurred with the new skill
-
Child devices were no longer correctly exposed to Alexa, so that many devices stopped working with the integration. Fibaro dual switches are a good example.
-
some customers, but not all, report that virtual sensors of many different kinds are no longer triggering Alexa routines. This is not all customers and has been erratic, with the devices not working then working again for a few hours or a few days, then failing again.
-
at least some customers who have multiple locations reported under the new skill the devices are all mixed together, which they were not under the old skill.
I’m not sure what’s happening with 3). Engineering is aware of 4) and is working on it and just said this morning that they now think they are zeroing in on a diagnosis. I haven’t seen any official responses to the multi location issue yet.
There is one other additional Longterm issue, which is the one that you just raised:
- on the old platform it was very easy for an individual customer to create a virtual device using the IDE and a copy/paste method. Or the virtual device creator smartapp. These were simulating hub-connected devices. Groovy is eventually going away (probably not until 2021) and smartthings has not Yet released the exact procedure for creating a custom DTH for a physical device that is connected directly via zigbee or Z wave.
And they haven’t said anything one way or the other as to whether there will be an easy way for an individual customer to create a virtual device on the new platform or whether it will require setting up a developer account and hosting a device as though it were a cloud cloud connection instead of a direct. We just don’t know. ( I have asked, twice, but did not receive an answer.)
So… I feel confident that engineering is doing its best to try to get the old platform virtual sensors Which use groovy DTHs to work with Alexa routines again Under the new Alexa skill. (4 above.)
I can’t tell whether any work is being done on three or five above.
They have said they are going to work on the duplicate devices issue, but I’m not clear on the technical approach to that.
I do not feel confident that they are going to work on changing the “all or nothing“ integration approach.
And I do not feel confident, to respond to your specific question, that we will be able to create virtual switches as easily once the IDE goes away. But since that’s not till 2021, I am putting off worrying about that too much right now.
Meanwhile, I did create a thread to list some of the alternatives to virtual switches that might be used if they just never come back. I’m not saying any of these are good choices, but they might be worth it for some people:
Sorry, I know that’s a long response with more questions than answers, but that’s my understanding at the moment.