This is an article about discoverability, not data going away or being deleted. This has been an issue for decades; some people with an axe to grind have sudden

This is an article about discoverability, not data going away or being deleted. This has been an issue for decades; some people with an axe to grind have suddenly found a reason to care now. The data exists but you have to know where to find it and this has never not been the case. Every administration does this and the underlying reasons are pretty banal. The infrastructure to deliver a copy of the data to everyone that wants it does not exist, not even close. Rationing is achieved by making data hard to find and gatekeeping. Some nominally public data is effectively too large to copy but there are very limited resources to work with it in place. Every time an administration changes out, this gets slightly rearranged and some data is prioritized and other de-prioritized but the underlying issue remains. There is also a lot of institutional reluctance to make public data actually available because of risk of "misuse", which mostly means doing analysis they don't want you to do with it. I got into a fight with USGS over access to a nominally public data set during the first Obama administration over this issue. Both Democrat and Republican administrations have a habit of wanting to control the narrative that can be spun with the raw data which makes it a bit of a political football. Which is to say, all of this is a side effect of practical data policy in the US government not being very good. The fact that there are hundreds of exabytes of this data makes it more than a policy challenge.