The transmission rate per-infected-capita has declined dramatically from the height of the epidemic and continues to drop:

http://www.microbiologybook.org/lecture/transmission-lg.gif

(Figures I can find indicate that the figures today are less than 2/3 the rate at the end of that graph circa 2005, and that while there are a few particular age subdemographics in which male-to-male sexual transmission rose slightly over the last decade total male-to-male transmission constantly declined.)

Annual figures from Russia indicate a massive decrease in spread in the early 2000s, and a major downward phase-shift in spread amongst infected people in the homosexual population circa 1996 when you compare the fraction that are infected via different routes with the total number of cases.

http://darussophile.com/2009/03/myth-of-russian-aids-apocalypse/

As near as I can tell the disease is now sub-replacement in the United States, with each person who gets it (a bit over half of all new infections now being male-to-male sexually transmitted) on average infecting less than one additional person over their expected lifespan, and this was true both just before and after antiretrovirals began massively extending life. That's a hell of a behavior change from the early days of the epidemic.