It's real, but not practical. The risk of death during a transplant like this is as high as 10%, because it involves completely knocking out the immune system.
Confounding the mortality stats is the fact that individuals undergoing treatment are very sick from the start. We might see better survival rates amongst comparatively healthy people.
In this case, you don't necessarily need a clean slate. "All" you need is a population of cells generating HIV resistant T-cells. AIDS manifests in individuals with extremely low T-cell counts. If a population of CD4 mutants could take hold within a larger population, it might be enough to prevent AIDS and bring viral titers low enough to eliminate transmission.
Nevertheless, I agree -- it's hardly practical, but it may be a beachhead.
Anybody remember that article? I'd like to find it again.
Here is a hypothetical example (because I don’t know much about HIV): Getting rid of frequent headaches (the hypothetical side-effect of managing a hypothetical otherwise deadly disease) with a cure that has a 1:5 chance of death doesn’t seem like a good deal to me.
HIV may not be a death sentence anymore, but it's still a serious disease with profound implications in one's lifestyle, and as far as I'm concerned, a cure can't come soon enough.
Admittedly this is pretty intense, but so are treatments for cancer. Let's hope this leads to improved and safer treatments in the future.