We watched the same thing from opposite ends: one of us inside a commercial warehouse's optimizer, deciding how queries run; the other building the distributed storage and infrastructure they run on. From both ends the conclusion was identical — engine choice had become the most expensive decision in the data stack, and it was being made once, by hand, and never revisited.
Open table formats quietly removed the technical reason a query must run where the data was loaded. Snowflake, Databricks, Google, and AWS now read the same tables — but nobody prices them against each other, query by query. Doing that well is not a dashboard problem; it's a query-optimization and distributed-systems problem: cost models that stay honest, translation that preserves semantics, verification that catches disagreement, and isolation that survives an auditor.
That happens to be the exact intersection of this team's two careers: years spent inside a commercial warehouse's query optimizer, and years building planet-scale distributed storage and infrastructure.
A quern is the oldest measuring machine there is — the hand-mill that grinds raw grain into exactly what it's worth. A kernos is a Greek ring-vessel with many cups attached to one ring: one vessel, many offerings. Quernos is both at once — one mill in front of many engines, weighing the worth of every query. The logo is the quern seen from above: the stone, its furrows at work, and the handle that turns it.
We're onboarding a small number of design partners — teams spending meaningfully on cloud warehouses who want the savings report and a seat at the table while the product hardens. Write to founders@quernos.com.