Four years after the Genomic Observatories Network was formally established as a collaboration between the Group on Earth Observations Biodiversity Observation Network and the Genomic Standards Consortium, we review the development of the network. Considering institutional infrastructure, we note the growing role of omic observation in active and increasingly interlinked marine networks, with examples such as EMBRC/ASSEMBLE, International Long Term Ecological Research Network, AtlantOS, National Association of Marine Labs, Smithsonian MarineGEO, and Partnership on Observation of the Global Oceans. We also note some key human elements essential to meeting the networks' goals, address how the community is evolving, and why performing seemingly simple tasks within a broadly distributed community presents significant challenges even among those who have agreed to use standards. From the perspectives above, we review lessons learned from use cases that leverage Genomic Observatories Network, such as the Autonomous Reef Monitoring Structures (ARMS), Ocean Sampling Day (OSD) and myOSD, which included experiences with citizen science. Looking forward, we survey 1) promising new technologies for in situ biological observation (e.g., cheap 3D printed omics samplers), 2) progress towards adoption of omics methods in marine policy and conservation programs, and 3) opportunities that a Genomic Observatory brings, alone or embedded in a network, to address novel scientific questions and support Essential Biodiversity Variables, Essential Ocean Variables, and indices such as the Ocean Health Index. Given the data intensive nature of omics investigation, we note emerging cyberinfrastructure solutions, such as the Genomic Observatories Metadatabase (GeOMe), an open-access repository for geographic and ecological metadata associated with biosamples, and predictive modeling efforts, such as those of the Island Digital Ecosystem Avatar (IDEA) Consortium. Finally, we explore the potential of Genomic Observatories as components of high-resolution calibration sites. Such observatories would provide super-contextualized "data trusts" for machine learning and artificial intelligence applications that draw on multi-omic observation.