Quoting my own post...in a post-necro attempt. Odd.
I see a lot of folks debating the validity of reporting out of various areas. I haven't seen anyone postulating that deaths or cases are being over reported, more under-reported for various political reasons. Which got me to thinking about what we can determine from the numbers that we do have...and they're not good.
If we go with the reported deaths in the US (even if under-counted), look at the graph of deaths by day (that one that shows the exponential growth). The area beneath the curve is the total number of deaths. The exponential growth gave a very sharp rise. While we've managed to flatten that curve, we haven't reached a plateau yet, much less a decline. When we do start declining, it's not going to decline at an exponential -- it's going to go down a LOT slower than that initial rise (barring the development of some miracle cure). That dramatically increases the area under the curve before we reach the end of this...and that area is the total number of deaths that we're looking at.
Taking nothing else into account (under-reporting, new variations/complications/etc), if we continue to keep the curve flattened so that we can start to trend downwards (i.e. continue the social distancing and other restrictions), we're likely looking at over 250k in deaths before this is done.