Folium can be very slow with a lot of data. One idea would be to only plot a random sampling of your points. If you are using pandas you can use plot_df = df.sample(1000) or something like that.

If you have specific data you want to show you could use a filter to filter out that data, so like plot_df = df.loc[df[<some column>] == <some value>, :]

Otherwise you could try geopandas instead of Folium, which is less computationally heavy. It won’t have all of the fancy features of Folium, though.