Last week, preliminary findings from a fifth serological study of 28,000 people in India's capital showed that 56% of residents already have antibodies, though a final report has not yet been published. The figures were higher in more crowded areas. Last summer, another survey by Mumbai's health department and a government think tank found that 57% of Mumbai slum-dwellers and 16% of people living in other areas had antibodies suggesting prior exposure to the coronavirus.
I think this will be a key reason, with many other factors at play. Immunity (natural or vaccine induced) is one of the factors that reduces the effective R number. You can't transmit the virus to someone who is immune. If 50% of the people you meet are immune, you will infect half the number of people you would have infected if nobody had immunity, in other words, the R number is halved.
It's a bit more complicated than that because infection with SARS-CoV-2 doesn't seem to give sterilising (total) immunity, ie you can be reinfected. However, in the same way that partial immunity from the vaccine reduces transmission, you would expect partial immunity after infection to reduce transmission too.
The problem is we believe immunity wanes over time so when that happens transmission will start to increase again.