zeeshan khan
Even while the software business attributed its demise to bad past decision-making, the collapse of Builder.ai has brought new attention to AI coding methods.
Supported by Microsoft, the national wealth fund of Qatar, and other venture capitalists, the British firm Builder.ai quickly became a near unicorn as its valuation drew close to $1 billion (£740 million). Although the Builder.ai team actually built the apps, the London company’s business strategy was to use AI tools to let users design and construct applications.
More than 500 million dollars in funds was pumped in by high-profile investors. But things weren’t going well at the beginning. The startup, formerly known as Engineer.ai, came under fire in 2019 when The Wall Street Journal disclosed that the majority of its coding was done by human engineers rather than artificial intelligence.
Read More: Microsoft-backed Builder.ai bankrupt after ‘AI’ powered by 700 Indian engineers
Although Builder.ai became more open about the human element, the business fell into financial difficulties. In February 2025, it named Manpreet Ratia as its new CEO, replacing founder Sachin Dev Duggal, whom it said was responsible for “transforming software development through AI-powered innovation.”
When funding suddenly ran out, it was Ratia’s responsibility to notify staff members on a call on May 20 that was covered by the Financial Times. that the company was declaring bankruptcy. Builder.ai apparently was unable to bounce back from previous difficulties and previous choices that imposed a significant strain on its financial position.”
Although startup failures are typical, even for well-known companies like Builder.ai, the company’s dependence on AI techniques to speed up coding may cause some consumers to pause and think.
There are a few examples of genuine utility among the flood of AI crap being produced by the tech sector. Generative AI can be a helpful programming assist in the field of coding, but when anticipated to behave like an engineer, they are frequently less than helpful.
“My new hobby: watching AI slowly drive Microsoft employees insane” is a humorous Reddit thread that links to multiple GitHub threads in the.NET runtime repository where people tolerate the GitHub Copilot coding agent’s numerous errors, many of which would make a junior developer blush. With the exception of coding instead of chess, everything feels a little like Mechanical Turk, and the human intervention is all too obvious.
“The amount of time they spend replying to a friggin LLM is just crazy,” one commentator stated. It’s depressing as well.
Microsoft CEO Satya Nadella boasted last month that artificial intelligence (AI) generated 30% of the code in some of the company’s projects. An observer is therefore compelled to believe that there is some passive assault going on here, as a developer has been instructed to employ the agent and will do so cheerfully. Satya Nadella isn’t afraid of layoffs, after all.
The issue brought to light by the merging demands in the.NET runtime repository and the failure of Builder.ai is that generative artificial intelligence (AI) tools are not a universal solution, despite the best efforts of tech giants looking for their next big growth opportunity, startups pitching the newest and greatest innovations, and management trying to reduce their budgets at the expense of engineers.
Although financial oversight along with unrealistic expectations were the main causes of Builder.ai’s death, the company was a favorite among generative AI coding companies and an excellent illustration of how a corporation might use the technology to streamline its operations.
Those who view generative AI as a substitute for junior developers should be cautious of this as it did not convince enough clients to pay it enough to remain sustainable. Technology still has a ways to go, as seen by the harmful Microsoft employees’ experience with the GitHub Copilot Agent. It might eventually outperform a lousy intern who knows how to use a search engine, but that day is not today.
