Skip to content
Companies

Code Metal

Verified translation of AI-generated and legacy code

Raised to date

$166M

across 4 Form D filings

Still open
$2M
Investors
12
Last filed
Jul 13, 2026

About

CodeMetal provides an enterprise software platform for verifiable code translation and optimization, focusing on AI-generated and legacy code. The company's technology uses neuro-symbolic verification and mathematical proofs to ensure code correctness, compliance, and safety, bridging the trust gap for deployment in mission-critical environments.

The company's platform stands out by using neuro-symbolic AI to mathematically prove the correctness of code. This allows it to bridge the trust gap associated with AI code generation, enabling safe deployment in highly regulated and mission-critical industries like defense and automotive.

Funding

$125M, first sold Mar 2026$123M sold$2M still openJun 2025DecJun 2026
Now
  • Option to acquireFiled Jul 13, 2026Form D
    None sold/ no target
  • EquityFiled Mar 12, 2026Form D
    $123M/ $125M
  • EquityFiled Nov 13, 2025Form D
    $36.5M/ $36.5M
  • EquityFiled Jul 10, 2025Form D
    $6.5M/ $6.5M

Product

Edge Workflow and Operations Management Platform

AI Code Verification Platform

An enterprise software platform designed for the automated translation, portability, and optimization of code across various programming languages and hardware environments. It verifies both AI-generated and legacy code to ensure it is correct, performant, and ready for production, particularly for edge devices in safety-critical settings.

  • Mathematically proves code correctness using neuro-symbolic verification
  • Supports automated code translation between different programming languages
  • Optimizes code for specialized hardware and edge computing environments

Built for

The primary customers are enterprises and government agencies operating in mission-critical sectors. This includes organizations in defense, automotive, semiconductor manufacturing, and other highly regulated industries.

People on file