Genomics by the Numbers: Costs Collapse, Cohorts Scale, Markets Mature

Sequencing costs and scale: the defining curve

In the Genomics sector, The genomics sector remains anchored by one of technology’s steepest cost curves. Over two decades, the cost to read a human genome has plunged from roughly $100 million to well under $1,000, a million-fold decline that outpaces Moore’s Law. That collapse is documented in long-running benchmarks maintained by the National Human Genome Research Institute, whose latest update shows costs continuing to slide as new high-throughput platforms come online NHGRI data.

Lower costs are translating directly into scale. The UK has assembled one of the largest human whole-genome resources to date, with 500,000 participants sequenced and linked to deep phenotypes—an asset designed to accelerate target discovery and population-level risk modeling UK Biobank release. In the United States, the National Institutes of Health’s All of Us Research Program has released nearly 250,000 whole genomes to qualified researchers, broadening representation in genomic datasets and creating a richer substrate for precision-medicine studies All of Us Research Program.

As the cost curve bends further, vendors are optimizing throughput and accuracy for distinct use cases. Short-read leaders are pushing per-genome consumable costs toward the low hundreds at production scale, while long-read platforms have increased output for structural variants and repeat expansions. The statistical implications are profound: large-N cohorts boost power for rare variant association, enable ancestry-aware polygenic scores, and shift the field from discovery to deployment.

Clinical adoption: from rare disease to population screening

Clinical genomics is moving from boutique to routine in select indications. Rare-disease diagnostics increasingly start with exome or genome sequencing, shortening diagnostic odysseys and informing care pathways. In oncology, comprehensive genomic profiling underpins targeted therapies and trial matching, while circulating tumor DNA is gaining traction for minimal residual disease monitoring and recurrence detection. The common thread is statistical maturity: larger variant databases, better priors, and richer phenotypes are improving signal-to-noise in real-world practice.