Tower Secures $6.4M for AI Data Engineering Platform

Jan de Vries · 2026-03-13

Let's talk about a problem that's keeping data engineers up at night. You know the feeling—AI can write code in minutes, but getting that code to actually *work* in production? That's where things get messy. Tower, a Berlin-based startup founded by former Snowflake engineers, just raised $6.4 million to solve exactly that. They're building what they call the "last-mile" platform for data engineering in the AI era. Think of it as the final stretch between having brilliant AI-generated ideas and actually running them on real company data. ### The Funding and The Vision The funding came through pre-Seed and Seed rounds, with Speedinvest leading the Seed and DIG Ventures taking the pre-Seed. They weren't alone—Flyer One Ventures, Roosh Ventures, and several angel investors from companies like Datadog and Motherduck joined in. That's a serious vote of confidence. Here's how CEO Serhii Sokolenko puts it: "With AI coding assistants accelerating development, the real challenge has shifted to production. Builders can create pipelines in minutes, but they need a platform to run them reliably on real company data." He's hitting on something we all sense. The internet's public data is outdated. Tower wants to power systems with information unique to your company. That's the game-changer. ### From Code to Production, Without the Headache Founded in 2024 by Sokolenko and CTO Brad Heller, Tower provides a single platform where humans and AI agents collaborate. It takes that AI-generated code and turns it into software systems businesses can actually depend on. It's designed for teams building scalable data and AI products. The platform brings together a few key things: - Python-native orchestration - Managed Apache Iceberg storage - Control plane APIs All in one multi-tenant system. The goal? To eliminate the need to stitch together a dozen different tools that never quite talk to each other right. Brad Heller shared this insight from their Snowflake days: "We saw how the next generation of engineers wanted a platform that truly combined data processing with AI. Today, developers are more productive with AI agents, but they still struggle with the same old operational problems." He's right. It's easier than ever to write functional code. But testing it, fixing issues, delivering to production, and operating it? That's still brutally difficult for humans—and nearly impossible for AI agents alone. That's the gap Tower is trying to fill. ### Why the "Last-Mile" Analogy Fits Besides transforming code into production-ready systems, Tower combines storage and processing compute on one platform. This gives data engineering teams serious analytics firepower. Their storage uses the Apache Iceberg open table format. This isn't just technical jargon—it means compatibility with Snowflake, Databricks, and all the major data engines. Customers own their data, and AI agents get fresh, company-specific information. This accuracy helps avoid those costly AI "hallucinations" we hear about. Florian Obst, a Principal at Speedinvest, nailed the value proposition: "Teams building vertical AI services need an analytics platform that integrates seamlessly without expensive legacy systems or complex cloud infrastructure. Serhii and Brad built a multi-tenant platform for fast integration and rapid iteration." ### The Traction Speaks Volumes The market seems to agree. Just a few months after launch in February, Tower reported some impressive numbers: - Over 200,000 runs of more than 30,000 unique applications - Their Python SDK hit 70,000 monthly downloads Builders worldwide, especially those creating Vertical AI services and SaaS products, are adopting Tower as their go-to platform. They're calling it the 'last mile' platform because it handles the final, crucial step—taking innovation from concept to reliable, daily operation. It's a simple but powerful idea. In an age where AI can generate solutions faster than we can implement them, Tower wants to be the bridge. They're not just another tool in the stack; they're trying to be the foundation that makes the entire AI-driven development process actually work for businesses. And with $6.4 million in new funding, they're gearing up to make that vision a reality for more teams.