The efficacy numbers you might be remembering could have been for the wild type, not the more infectious delta variant. A higher infectiousness also requires a far higher fraction of the population to be vaccinated to reduce spread (rule of thumb is that a fraction of
1 - 1/R_0 would be needed for a given value of R_0 iirc).
You also have to keep in mind that the incidences for vaccinated and unvaccinated are vastly different. Several German federal states record these values and the incidence among the unvaccinated is usually ~10x as high as for the vaccinated, so a small fraction of the population can be responsible for a good deal of the infections. I shall provide a link for the latter in a moment.
Example for Saxony, one of the more affected federal states at the moment. Ratio of incidences for unvaccinated:vaccinated is ~19:1, see:
https://www.coronavirus.sachsen.de/infektionsfaelle-in-sachs...
Example for Bavaria, also heavily affected. Ratio of incidences for unvaccinated:vaccinated is ~9.7:1. Source:
https://www.lgl.bayern.de/gesundheit/infektionsschutz/infekt...
Baden-Württemberg, taken from an article as I couldn't find the original source. Ratio of incidences unvaccinated:vaccinated is ~18.4:1 there:
https://www.swp.de/panorama/intensivstation-geimpfte-corona-...
Keep in mind that you would need to weigh that according to the vaccination levels if you want to calculate somethign with them but you should find the current levels on the same pages, I think.