While everyone's talking about AI coding, the real money might be in fixing what gets built. Momentic, an AI-powered software testing startup, just closed a $15 million Series A led by Standard Capital with Dropbox Ventures joining the round. The funding signals growing investor confidence that AI can solve one of development's most persistent headaches - making sure code actually works before it ships to users.
The unsexy side of software development just got a major cash injection. Momentic announced Monday it raised $15 million in Series A funding led by Standard Capital, with Dropbox Ventures participating alongside existing investors Y Combinator, FCVC, Transpose Platform and Karman Ventures. The round builds on a $3.7 million seed announced in March, bringing total funding to nearly $19 million.
While product demos grab headlines, co-founder Wei-Wei Wu knows the real work happens behind the scenes. "Testing has been the biggest pain point for every team I've ever worked with," Wu told TechCrunch. His solution? Let AI handle the tedious but critical work of quality assurance that keeps software running smoothly.
Momentic's pitch is deceptively simple - developers describe their critical user flows in plain English, and the company's AI automatically creates and runs the tests. "We help our customers make sure their product works," Wu explained. "They can describe their critical user flows in plain English and our AI will automate it." It's a stark contrast to existing open-source frameworks like Playwright and Selenium, which offer granular control but require significant technical expertise to implement effectively.
The approach is already resonating with major players. Momentic counts 2,600 users across companies including Notion, Xero, Bilt, Webflow, and Retool. Wu stayed tight-lipped about revenue figures, but said growth metrics were strong enough to convince investors. The scale becomes clear when you look at the numbers - Wu estimates Momentic automated over 200 million test steps in the past month alone, a volume that would be virtually impossible with traditional manual testing approaches.
Wu and co-founder Jeff An aren't newcomers to developer pain points. Both cut their teeth on developer tooling at companies like Qualtrics and WeWork, with Wu particularly proud of his contributions to open-source Node.js. That experience gave them front-row seats to watch teams struggle with code verification across different environments and use cases.
But Momentic faces an unusual competitive threat - the very foundation models that power its AI could eventually make the company obsolete. Both OpenAI and Anthropic now offer tutorials on agentic testing, building on their models' improving computer use capabilities. As these foundation models become more sophisticated, the moat around enterprise SaaS companies like Momentic could narrow significantly.
Wu isn't worried yet. The new funding will go toward expanding the team and building more sophisticated features. The company launched mobile testing support in August and plans to develop advanced test-case management tools once more engineers come aboard. Wu sees the rise of automated coding as a tailwind rather than a threat - more AI-generated apps means more demand for quality assurance.
"All of these apps need testing," Wu said. "They care about quality, and we're going to provide it for them." It's a bet that even as AI gets better at writing code, human oversight of that code will remain essential. The question is whether specialized tools like Momentic can stay ahead of the foundation models long enough to build a defensible business.
Momentic's $15 million raise highlights a crucial tension in AI development - while everyone focuses on generating code, someone still needs to make sure it works. The company's bet that AI can automate quality assurance at scale is paying off with major clients and impressive testing volumes. But with OpenAI and Anthropic building their own testing capabilities, Momentic faces a race against time to establish itself before foundation models potentially eliminate the need for specialized tools. For now, the growing complexity of software and Wu's vision of an app-heavy future suggest there's plenty of testing work to go around.
