MotherDuck, the serverless data-warehouse platform, has purchased German startup Tower in a deal announced August 25, 2026, bringing a managed Python-pipeline runtime and AI-driven control interface in-house. The transaction, MotherDuck's first acquisition in its four-year history, did not disclose financial terms. The deal accelerates MotherDuck's push into AI agents that build and operate data pipelines by eliminating reliance on a third-party component.
Tower, founded in late 2024 by former Snowflake engineers Serhii Sokolenko and Brad Heller, had already become MotherDuck's largest customer after its runtime was adopted to power the "Flights" feature—a general-purpose Python execution layer for AI agents. The acquisition folds Tower's technology and engineering team into MotherDuck's product organization, giving the company direct control over sandboxing, credential management, and scheduling capabilities that underpin its AI-driven pipeline automation. MotherDuck currently operates two sandboxed runtimes: Flights, now backed by Tower's technology, and Ducklings, its native serverless DuckDB instances. The plan is to eventually merge these environments, combining Ducklings' instant startup with Tower's robust job scheduling and credential isolation, while existing Tower customers will be migrated to MotherDuck's Flights offering.
According to CEO Jordan Tigani, the company could have built a similar layer internally, but the time-to-market advantage and the proven reliability of Tower's platform tipped the balance toward acquisition. The report finds that MotherDuck's core proposition revolves around a hybrid execution model that lets queries run locally on a laptop, in the cloud, or across both, leveraging the open-source DuckDB engine. Recent product extensions, such as the MCP-based AI agents and the Flights runtime, require a reliable, secure environment to execute code generated by large language models, and Tower's managed runtime supplies that missing foundation.
The acquisition reflects a broader industry shift where data-warehouse vendors are moving beyond query-only interfaces toward AI-enabled data engineering, the report states. Competitors such as Databricks and Snowflake have introduced their own code-generation agents, but MotherDuck's approach of embedding a managed runtime directly into its serverless architecture differentiates it by offering end-to-end execution without relying on external compute services. By owning the runtime, MotherDuck can more tightly align product roadmaps, pricing, and security guarantees, positioning itself as a one-stop shop for AI-driven data pipelines. The report notes that competitors that continue to rely on third-party runtimes may face longer development cycles and higher integration overhead, giving MotherDuck a tactical edge in the emerging AI-agent market.
For MotherDuck, the report concludes, the acquisition eliminates a critical dependency and gives the company full ownership of the execution layer that powers its AI agents, translating into faster feature rollouts, tighter security postures, and the ability to bundle runtime costs into its subscription pricing, potentially improving net revenue retention. Tower's customers gain a clear migration path to a platform that promises deeper integration and a unified billing experience, with the trade-off being access to a more tightly coupled stack that can deliver lower latency and more predictable performance for AI-generated pipelines. In the longer term, the combined entity could command higher ARR multiples as investors reward the added moat of owning both the data warehouse and the pipeline execution foundation. The vertical integration strategy may force rivals to choose between building proprietary runtimes or accepting slower innovation cycles, reshaping competitive dynamics across the data-infrastructure market.