Code Metal raises $125M to modernize defense code with AI

Boston startup Code Metal has announced it raised $125 million in funding. The news says something about more than one company's ambitions: it points to a broad shift in how the defense industry approaches its technological debt. For decades, military contractors have run software systems written in COBOL, Fortran and other languages of the last century, keeping them alive through the efforts of a few specialists whose average age keeps rising. Now AI is taking on a task people have put off for too long.

Legacy code in the defense sector is one of the most underestimated technological threats of our time. Critical control systems, logistics platforms and guidance software often run on code that is thirty, forty or even fifty years old. Attempts to rewrite it by hand traditionally end the same way: projects drag on for years, costs overshoot budgets many times over, and new bugs reproduce or even worsen the original's problems. This point of failure is exactly where Code Metal positions itself.

The key difference in the company's approach is not translation speed but verification. Most tools built on large language models can rewrite code from one language to another at impressive speed, but they cannot guarantee that the result behaves identically to the original. In a consumer app, a small discrepancy means a bug report. In a missile control system or a target recognition algorithm, it means catastrophe. Code Metal claims its AI system does not merely translate code but mathematically proves that the new code is equivalent to the old: every branch of logic, every boundary condition and every side effect goes through formal verification.

This field, known as formal verification, has existed in academia for a long time, but traditionally demanded enormous manual effort and deep mathematical expertise. Code Metal is betting that modern AI models make it possible, for the first time, to automate the process at industrial scale. If it really works as the company says, it would be a qualitative change in what is possible when migrating legacy systems.

The $125 million round is a weighty signal to the market. The defense technology sector is in a genuine investment boom after several years in which Silicon Valley ostentatiously distanced itself from military contracts. The success of Anduril, Shield AI and other defense-tech startups has reshaped how the industry is perceived: defense technology is once again considered an attractive venture category. Code Metal rides this wave with a product that addresses a concrete, measurable pain, not an abstract promise of "security" or "efficiency."

For defense contractors themselves, the stakes are extremely high. The US Department of Defense has spent years trying to modernize its technology base through programs like JEDI and its successors, but the infrastructure layer, the software itself in outdated languages, remains a painful bottleneck. Finding a Fortran programmer who can make sense of forty-year-old code without documentation gets harder every year. Automated migration with a guarantee of correctness is not just a convenience; it is a matter of long-term combat readiness.

Code Metal also faces obvious skeptics. Formal verification of complex systems is a problem computer science has not fully solved, even outside the AI context. Real defense code often contains intentional or accidental dependencies on specific behavior of hardware, the operating system or the compiler, nuances that are extremely hard for any automated tool to catch. Proving that the technology works on live systems, not demo projects, is a task the startup still has to accomplish publicly and convincingly.

Still, the very appearance of Code Metal with serious funding marks an important moment: the industry has finally recognized that AI is ready to take on not only writing new code but also unwinding decades of accumulated technological debt. If the bet proves right, the next few years will show how the very idea of maintaining critical software systems is changing, not only in defense but in energy, aviation and finance. Legacy code is everywhere. The problem is universal. Code Metal simply started with the most demanding customer.