https://www.cdc.gov/hiv/group/msm/index.html
Canada -> 53% of new HIV infections are MSM http://www.catie.ca/en/fact-sheets/epidemiology/epidemiology...
Australia - > 68% of new HIV infections are MSM http://www.hivmediaguide.org.au/hiv-in-australia/hiv-statist...
UK -> 54% of new HIV infections are MSM http://www.catie.ca/en/fact-sheets/epidemiology/epidemiology...
Globally the situation is much different and about the same number of men as women have HIV and drugs do not appear to be a big factor in to the equation.
It is interesting how the data has changed over the years, for a long time gay men were not the largest group but maybe with the increased effectiveness of anti-virals combined with decreased mortality the balance has shifted substantially once again.
In the beginning of the AIDS epidemic it was basically a disease afflicting predominantly gay people and only rarely heterosexuals. The two groups being somewhat socially isolated was an important factor in this. Then the balance shifted as more and more heterosexual people got infected and was made worse by people thinking that HIV would pass them by because they were not gay.
So now we have come full circle.
Still, 32% (at best) are pretty lousy odds so better be very careful.
That's not how statistics work.
MSM and IV drug users are a small percent of the overall population. If you looked at the rate of HIV cases and divided by the respective population, you'd see that MSM and IV drug users are at a far greater risk of contracting HIV than heterosexuals (by magnitudes of order).
However, that certainly doesn't infer that heterosexual can't get AIDS and that they shouldn't seek to avoid it through appropriate measures.
For the mainstream population that is a much lower risk than for a smaller group that has a roughly equal incidence in absolute numbers.
But that risk is still substantial, and you do not want to end up as an entry in the tally of that 32%, which is a substantial number of individuals, if approximately 4% is LGBT then maybe 50% of those (so 2% of the population) takes 68% of the cases, leaving 98% to divide the remaining 32%.
There are many lotteries with far fewer winners.