Kodesage raises $6.6M to drag enterprise legacy software into the AI era, without it leaving the building
What's the deal? London and Budapest-based Kodesage has raised $6.6M in seed funding to help banks, insurers, and other regulated enterprises modernise decades-old legacy software using AI — all without data ever leaving the customer's environment.
The round is led by VentureFriends, with pre-seed lead Portfolion returning. Angel investors include Christian Szegedy, a co-founder of Elon Musk's xAI, and Mario Götze, the footballer who scored Germany's 2014 World Cup-winning goal. It follows a roughly €2.3M pre-seed raised in early 2025.
Founded in 2024 by Gergely Dombi, Miklos Szurdi, and Gyorgy Szilagyi, the startup targets legacy systems written in COBOL, PL/SQL, PowerBuilder, and Oracle Forms — software that still runs core processes at major institutions but is poorly documented and understood by fewer and fewer people. Dombi and Szurdi previously built a 300-plus-person consultancy doing this work manually. Szilagyi co-founded Tresorit, the encrypted-storage company Swiss Post acquired in 2021.
Kodesage's platform performs automated deep discovery of legacy codebases, extracts buried business logic from stored procedures and database schemas, and turns it into documentation that human teams and AI agents can read. From there it supports code conversion, automated test generation, and AI-assisted production support.
Why now? The engineers who wrote these systems are retiring, and the knowledge is leaving with them. Meanwhile, cloud-based AI coding tools can't access the most sensitive enterprise data — compliance rules keep it behind the firewall. That gap is Kodesage's opening.
Its platform runs entirely on-premise, in a virtual private cloud, or fully air-gapped. Source code and database content never leave the organisation's control. This also decouples costs from per-token cloud pricing, offering more predictable bills.
What could go wrong? Kodesage's performance claims — up to 3x faster migrations and 80% less documentation effort — are company figures, not independently audited. Chief executive Dombi describes the legacy modernisation problem as a "multi-trillion-dollar drag," but that's a sizing claim rather than a measured one.
The startup's stated vision of "self-healing enterprise applications" — where AI agents prepare, test, and implement fixes while engineers only review — is framed as a direction of travel, not a shipping product. Getting from automated documentation to autonomous code fixes is a big leap.
The signal: VentureFriends, the Athens-headquartered fund leading this round, has built a portfolio heavily weighted toward fintech and enterprise software — making a legacy-modernisation play in regulated finance a logical extension of that thesis. The round also underscores growing investor appetite for AI infrastructure that operates inside the enterprise perimeter rather than through cloud APIs, a deployment model that could prove decisive in sectors where data sovereignty rules effectively block the dominant coding assistants. For Kodesage, the real test is whether an on-premise-first approach can scale as efficiently as cloud-native competitors while the founders' consulting roots give them the domain credibility to land early bank contracts.
Read more: thenextweb.com
