This model was put out by a group at University of Washington. It assumes social distancing continuing throughout May. Since we may not actually do that, the real curve may end up looking a bit different on the back side.
The low end scenario has around 40,000 US deaths by early August. You can see the high end is over 140,000. Again this is with social distancing continuing throughout the period and obviously assuming no effective therapeutic comes along in the meantime. If we went completely back to life as normal then the numbers would be way uglier than this.
In the nearer term, the low end scenario still has 1,000+ deaths per day in the US from April 5-19.