Clustering of large vegetation dataset with NMDS and k-means (R package vegan)

Question

I am working with a very large dataset of vegetation cover data (300,000 plots x 10,000 species) that I want to get clustered in vegetation types. My normal workflow would be to