AI coding created a new bottleneck: $1.5B says validation is it
Two funding rounds landed in the same week, pointing at the same problem. AI code review platform CodeRabbit reached a $1.5 billion valuation after its latest round. Software testing infrastructure company Blacksmith raised $45 million at a valuation approaching $550 million. Both are betting that AI-assisted coding has outpaced the tooling that checks what it produces.
The bottleneck has moved
For years, writing and shipping code was the slow part. AI has compressed that. The constraint now sits downstream: reviewing, validating, and testing what the AI generated.
“AI has done two things: It’s increased the rate of code generation, but it’s also exposed the bottleneck of checking, validating and sanitizing that code. We’re moving away from developers being the bottleneck to the testers being the bottleneck.” — Richard Simon, CTO, T-Systems International
The shift is structural, not just a tooling gap. Engineers who write code themselves carry context about why decisions were made. With AI-generated code, that reasoning is absent even when the output looks clean. That makes validation harder than running a test suite.
The closed-loop risk in AI-generated tests
The problem compounds when AI also writes the tests. Tommy Tran, a software engineer at Meta working on AI and machine-learning infrastructure, puts it directly: if your tests were generated by the same system that wrote the code, you risk testing whether the AI agrees with itself rather than whether the implementation meets the actual requirements.
Balaji Srinivasan, senior director of engineering at LinkedIn, adds that conventional coverage metrics (the percentage of code exercised by tests) become less useful when AI can inflate both numbers cheaply. What matters is whether the tests are validating meaningful behavior, not just whether coverage percentages are high.
What this means for your stack
CodeRabbit automates code review. Blacksmith handles continuous integration and testing infrastructure. The investment thesis behind both is that organizations with weak automated testing will hit a ceiling on how much they can benefit from AI coding tools, because a human still has to validate every change before it ships.
Tran notes that building these validation systems internally requires ongoing maintenance as AI models, security risks, and dev environments change. Specialized vendors can spread that cost across customers. The tradeoff is that buying the tools does not transfer responsibility for software quality. Requirements, risk tolerance, and ship decisions stay with the team.
If your testing coverage is thin today, faster code generation just means faster accumulation of unreviewed changes. That is the problem both of these companies just raised hundreds of millions to solve.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
