Metagenomics emerged in the late 1990s as a tool for accessing and studying the collective microbial genetic material in the environment. The advent of the technology generated great excitement, as it has provided new opportunities and technologies for studying the wealth of microbial genetic diversity in the environment. Metagenomics has been widely predicted to access new dimensions of protein sequence space. A decade on, we review how far we have actually moved into new sequence space (and other aspects of protein space) using metagenomic tools. While several novel enzyme activities and protein structures have been identified through metagenomic strategies, the greatest advancement has been made in the isolation of novel protein sequences, some of which have no close relatives, form deeply branched lineages and even represent novel families. This is particularly true for glycosyl hydrolases and lipase/esterases, despite the fact that these activities are frequently screened for in metagenomic studies. However, there is much room for improvement in the methods employed and they will need to be addressed so that access to novel biocatalytic activities can be widened.