Qodo secures $70M to expand code verification as AI-driven coding grows

As AI-powered coding tools continue to produce billions of lines of code each month, a new challenge is becoming increasingly apparent: ensuring that the software actually functions as expected. Qodo, a company focused on building AI agents for code review, testing, and governance, is positioning itself around the idea that verification will play a central role in the next stage of software development.

The New York-based startup has secured $70 million in a Series B funding round led by Qumra Capital, bringing its total funding to $120 million. The round also included participation from Maor Ventures, Phoenix Venture Partners, S Ventures, Square Peg, Susa Ventures, TLV Partners, Vine Ventures, as well as individual investors such as Peter Welinder of OpenAI and Clara Shih of Meta.

Qodo is focused on building a layer that improves trust in AI-generated code as enterprises accelerate their adoption of tools such as OpenClaw and Claude Code. Many organisations are realising that faster code generation does not automatically translate into secure, reliable, or production-ready software.

Unlike many AI-based review tools that focus on identifying what has changed in a codebase, Qodo emphasises understanding how those changes affect entire systems. Its approach takes into account organisational standards, historical development context, and varying levels of risk tolerance, helping teams manage AI-generated code with greater confidence.

The company was founded in 2022 by Itamar Friedman, who previously co-founded Visualead and later led machine vision efforts at Alibaba after Alibaba acquired the startup. Friedman has said that two experiences shaped his thinking: his time at Mellanox, which Nvidia later acquired, and his work building Visualead.

At Mellanox, he focused on automating hardware verification using machine learning, which led him to conclude that generating systems and verifying them require fundamentally different tools and approaches. Later, while working at Alibaba’s Damo Academy, he observed the rapid evolution of AI systems capable of reasoning over natural language. By 2021 and 2022, just before the emergence of GPT-3.5, it became clear to him that AI would generate a significant portion of global content, especially software, reinforcing the need for dedicated verification systems.

Recent data support this concern. A survey found that while 95% of developers report not fully trusting AI-generated code, only 48% consistently review it before committing changes. This highlights a gap between awareness of potential risks and actual development practices.

Friedman argues that large language models alone are not sufficient to ensure software quality and governance. Code quality often depends on internal standards, previous engineering decisions, and what he described as “tribal knowledge” within organisations. Even highly capable engineers may struggle to review code effectively in a new environment without that context, and the same limitation applies to AI systems.

While companies like OpenAI and Anthropic are shaping the broader AI ecosystem, including coding tools, they are primarily focused on building features rather than comprehensive end-to-end verification platforms. Although several startups are exploring this space, many remain early-stage and have not yet achieved widespread enterprise adoption.

To differentiate itself, Qodo is focusing heavily on performance. The company recently ranked first on Martian’s Code Review Bench with a score of 64.3%, outperforming its nearest competitor by more than 10 points and surpassing Claude Code Review by over 25 points. The benchmark highlights the system’s ability to identify complex logic errors and cross-file issues without overwhelming developers with unnecessary alerts.

Over the past month, the startup has introduced Qodo 2.0, a multi-agent code review system that currently leads industry benchmarks. It also launched tools that learn and adapt to each organisation’s definition of code quality, further tailoring its capabilities to enterprise needs.

The company is already working with major organisations, including Nvidia, Walmart, Red Hat, Intuit, and Texas Instruments, as well as fast-growing technology firms such as Monday.com and JFrog.

According to Friedman, the evolution of AI in software development has moved through several defining moments, from tools like Copilot to conversational systems like ChatGPT and now toward full task automation. He believes the industry is now entering a new phase, shifting from stateless AI systems toward stateful systems that incorporate deeper contextual understanding — a transition he describes as moving from intelligence to what he calls “artificial wisdom.”