Tobiko Data was founded by data infrastructure engineers from Airbnb, Apple, Google, and Netflix who built and operated transformation systems at massive scale. They launched publicly in June 2024 with a clear thesis: the dominant data transformation workflow wastes compute, breaks easily, and makes it dangerous to ship changes. Tobiko's answer is SQLMesh, an open-source transformation framework, and SQLGlot, a widely used SQL parser and transpiler that understands dozens of dialects.
SQLMesh introduces virtual data environments that let engineers build and preview transformation changes against real data without duplicating expensive compute or polluting production. It computes column-level lineage automatically, understands which models actually changed, and only recomputes what is necessary, which can dramatically reduce warehouse costs. Built-in testing, automatic change categorization, and previews make shipping transformation changes far safer than the test-in-production patterns common elsewhere.
SQLGlot, Tobiko's other open-source project, is a no-dependency Python SQL parser, transpiler, and optimizer that can translate queries between dialects like Snowflake, BigQuery, Spark, DuckDB, and Postgres. It underpins SQLMesh's cross-engine intelligence and is widely adopted across the data ecosystem in its own right. On top of these open-source foundations, Tobiko Cloud provides managed orchestration, scheduling, and observability for teams that want a fully supported experience.
Tobiko Data raised a $17.3 million Series A in June 2024 led by Theory Ventures, with participation from Unusual Ventures, 20Sales, and angels including Fivetran CEO George Fraser, Census CEO Boris Jabes, and MotherDuck CEO Jordan Tigani, bringing total funding to roughly $21.8 million including a prior $4.5 million seed. Tomasz Tunguz of Theory Ventures joined the board. Tobiko competes directly with dbt, positioning SQLMesh as a more efficient, correctness-focused alternative.