Code Metal Raises $125 Million to Rewrite Legacy Defense Software With AI
The Boston startup uses AI to translate and verify legacy code for defense contractors, arguing modernisation cannot come at the cost of new bugs
The funding positions Code Metal as one of the more heavily capitalised startups in the emerging 'vibe coding' space, though with a distinctly enterprise and government focus rather than consumer-facing tools.
The defense industry has long struggled with massive codebases written in languages like COBOL, Ada, and Fortran that are increasingly difficult to maintain as the engineers who wrote them retire. Code Metal's approach uses AI to translate these codebases into modern languages while providing formal verification that the translated code behaves identically to the original.
The company argues this verification step is what separates it from general-purpose AI coding tools, which can introduce subtle bugs that would be unacceptable in mission-critical defense systems.
Analysis
Why This Matters
Defense modernisation is a multi-billion dollar problem, and the US government has been increasingly open to AI-assisted solutions. A $125 million raise signals serious investor confidence in the defense-AI intersection.
Background
Legacy code translation has been one of the earliest practical applications of large language models. The defense sector is a natural fit given its enormous technical debt and deep pockets.
What to Watch
Whether Code Metal's verification claims hold up under scrutiny. The defense procurement cycle is notoriously slow, so revenue growth will be the real signal.