Today we are thrilled to share the news of our seed and follow-on Series A investments in Tobiko Data, which is re-imagining best-in-class transformation for data teams everywhere.

‍

We believe early-stage investing, especially at the formation stage, requires partnering with founders who have a strong vision and deep insights, which are oftentimes rooted in lived experience with the fundamental problem to be solved. Tobiko’s founders — Toby, Iaroslav, and Tyson — epitomize how that background can translate into a powerful advantage when bringing a new product to market. 

‍

Toby and Iaroslav were key technical leaders for Airbnb’s metrics platform, Netflix’s experimentation platform, and Apple’s AI/ML infrastructure, where their colleagues had raved about their brilliant data engineering work. Those systems, and similar ones elsewhere, are more critical than ever in helping everyone from analysts to executives leverage data to understand what is happening within their businesses. It also turns out the data pipelines that power them can be extremely brittle and inefficient in how they process data, especially when it comes to transformation. Changes to queries and tables can break things downstream. Models may be fully rebuilt, causing warehouse compute costs to spiral and scalability issues. Engineers can’t just rely on SQL but have to work with Jinja as well. The list goes on.

‍

In short, Toby, Iaroslav, and Tyson recognized that transformation, the stage between ingest and analytics, is essential to data productivity within every organization and is also the weakest link in the modern data stack today, despite a proliferation of tools. For some time, we had also been hearing that data teams were frustrated by the limitations of existing transformation options from dbt and others. When the Tobiko team outlined their idea of starting with an open-source project that would focus on SQL as its “lingua franca”, be interoperable across multiple data engines, and drive innovations in data pipelines, it immediately resonated with us. Our team also believed that a new, unique approach to data transformation was sorely needed. And among other things, we have a long track record in backing open-source companies.

‍

‍

‍

Since we led Tobiko’s seed investment in the fall of 2022, the team has executed at a torrid pace. Toby and Iaroslav have assembled a world-class team of engineers. It doesn’t take long to find commentary about how fast they ship, how responsive they are, and how knowledgeable they are. The community around Tobiko’s open-source project, SQLMesh, has quickly grown to thousands. There is no shortage of blog posts on why data teams should move from dbt to SQLMesh. Companies including Fivetran, Harness, Textio, Pipe, Dreamhaven, Zitcha, and many others have made SQLMesh the foundation of their transformation efforts. The project has incorporated industry-changing features such as column-level lineage, unit testing, and virtual data environments. With today’s launch of their new enterprise product, there is a lot more in store to help data teams build and automate robust data pipelines to power their most critical business needs.

‍
We are excited to welcome Tomasz Tunguz from Theory Ventures and the CEOs of Fivetran (George Fraser), Census (Boris Jabes), and MotherDuck (Jordan Tigani) in supporting Tobiko in their journey to build a truly transformational company (pun intended). We believe this is just the beginning of how Tobiko will reshape data stacks everywhere for years to come. Onward!