Tower raises €5.5 million to operationalise AI-generated software

91% of engineering organisations now use at least one AI coding assistant, according to the DX AI-assisted Engineering Q4 Impact Report 2025. Yet adoption has not translated uniformly into productivity gains. A randomised controlled trial by METR (Becker et al., 2025), in which 16 experienced developers completed 246 real-world tasks, found that those using AI tools took 19% longer to complete them than those working without - despite estimating they had been 20% faster. The primary challenge, the study suggests, has moved from code generation to reliable deployment.

Tower, a Berlin-based data infrastructure company founded by Serhii Sokolenko (CEO) and Brad Heller (CTO) - both former Snowflake engineers - addresses precisely that last mile. The platform provides flexible compute and analytical storage management in a single environment, giving data engineering teams the infrastructure to turn AI-generated pipelines and agents into production-ready systems powered by live, company-specific data.

Built on the Apache Iceberg open table format, it ensures compatibility with Snowflake, Spark and the broader data engine ecosystem without locking organisations into a single provider. Teams can process workloads on self-hosted runners, keeping sensitive data within their own environment, or on Tower's fully managed serverless infrastructure, with unified observability across both.

Recently, the company raised €5.5 million across pre-seed and seed rounds, with DIG Ventures leading the pre-seed and Speedinvest leading the seed, alongside Flyer One Ventures, Roosh Ventures, Celero Ventures and Angel Invest, and angels including Jordan Tigani (MotherDuck’s CEO), Olivier Pomel (Datadog’s CEO) and Maik Taro Wehmeyer (Taktile’s CEO).

Sources: Tower | DX AI-assisted Engineering Q4 Impact Report 2025 | METR, Becker et al. (2025)

Founders: Serhii Sokolenko, Brad Heller