Qodo raises $70M, betting on code verification in the age of AI development

While competitors race on code generation speed, Qodo is betting on something else entirely: confidence that generated code actually works. The company has raised $70M to tackle a problem many players in the AI development market prefer to stay quiet about: the quality of code written by AI.

The AI development industry is going through a paradoxical moment. AI-based tools, from GitHub Copilot to Cursor and Devin, generate code faster than ever. Developers accept hundreds of suggestions a day, and teams have cut the time they spend on routine code several-fold. But speed has brought a new problem: the volume of code that needs checking is growing faster than the ability to verify it. According to various studies, developers who use AI assistants heavily spend a significant share of their working time not writing code but reviewing automatically generated code.

This is the niche Qodo sees for itself. The company, formerly known as CodiumAI, has spent several years building AI-powered code testing tools. In this round, it sharpens its positioning: code generation is a solved problem, verification is not.

The logic behind the thesis is clear. When a developer writes code by hand, they understand what every line does. When AI generates a function of several dozen lines from a single prompt, the developer often takes it on trust, especially if the code looks convincing and passes basic tests. Problems surface later: in edge cases, at scale, in production. This is not a hypothetical threat. It is a real shift in the structure of technical debt, and many teams are already seeing it.

Qodo offers several answers to this problem. The key product is automatic test generation that does not merely cover the happy path but actively hunts for edge cases and potentially problematic behavior. The second component is code analysis at the level of business logic, not just syntax: the system tries to understand the developer's intent and check whether the implementation matches it. The third is integration into CI/CD pipelines, so verification runs automatically on every commit.

$70M is a serious sum, especially as the venture market cools. It suggests investors believe in the company's thesis. As AI agents start writing more and more code with minimal human involvement, quality control becomes critically important and commercially attractive.

Competition in the niche is heavy. Standard static analysis tools have existed for decades. Major players such as Snyk, SonarQube and Veracode have long been checking code, and GitHub is adding analysis features directly to its product. But Qodo argues that traditional approaches cannot cope with AI-written code, which differs in structure, style and error distribution. This is not just a marketing narrative; a real technical problem lies behind it.

Qodo is betting that verifying AI code requires specialized tools trained on the specifics of exactly that code, and that as AI agents begin writing entire systems rather than individual functions, demand for such tools will only grow.

The $70M will let the company scale product development and move into the enterprise market. The emerging segment of AI code quality checking has no recognized leader yet, and Qodo clearly intends to claim that position. The next 12-18 months will show how well-founded that claim is.