Harjot Gill was running Fluxninja, an observation start that he founded several years after the sale of his first start Net in Nutanix In 2018, when he noticed a strange trend.
“We had a team of remote engineers who were starting to adopt AI Code Generation at Github Copilot,” Gill told TechCrunch. “We have seen that adoption is happening and it was very clear to me that as a result of a second order, it would cause congestion in the revision of the code.”
At the beginning of 2023, Gill started Coderabbit, a Code review platform and acquired Flexninja.
Gill’s prediction has come true: Developers are now using AI encoding assistants to create code, but the output is often buggy, forcing engineers to spend a lot of time on corrections.
Coderabbit can help catch some of the errors. The company is increasing by 20% a month and now makes more than $ 15 million in annual recurrent revenue (ARR), according to Gill.
Investors find the development of starting fascinating. On Tuesday, Coderabbit announced that it has increased a $ 60 million B -raising series, rating the company to $ 550 million. The round, which brought the total funding of the start to $ 88 million, was led by Scale Venture Partners with Nvidures, Nvidia’s ARM Capital Capital and returning investors, including CRV.
Coderabbit helps companies such as Chegg, Groupon and Mercury, along with over 8,000 other businesses, save time on the famous frustrating work of the Code Review, which has become even more time consuming with the rise of the Code produced by AI.
TechCrunch event
Francisco
|
27-29 October 2025
Since Coderabbit understands a company code, it can detect errors and provide feedback, acting as a partner, Gill said. He added that companies using Coderabbit can reduce the number of people working in the code by half.
As with most areas of AI, Coderabbit has competition. Starting opponents include graphite, which secured a $ 52 million B series, led by ACCEL earlier this year and Greptile, which we mentioned are in talks for a number of $ 30 million with the tour with the benchmark.
While leading AI coding assistants such as the code and the Antropic runner, they also offer AI -powered code capabilities, Gill bets that customers will prefer an autonomous bid in the long run. “Coderabbit is much more complete in terms of depth and technical width from federal solutions,” he said.
Either its prediction will prove to be right, it remains to be seen. But for the time being, thousands of developers are clearly happy to pay Coderabbit $ 30 a month.
Even with the growing popularity of AI Code Review Tools such as Coderabbit, AI solutions cannot yet fully trust to correct the errors and the “useless” code written by AI. The unreliability of the code created by AI has caused a new corporate role: The Vibe Code Cleaning Specialist.
