In the golden age of AI-powered storytelling, Builder.ai had a pitch too good to ignore: make software “as easy as ordering pizza.” The startup, founded in 2016, claimed it could democratise software development by letting non-engineers build complex apps through a supposedly AI-driven platform. Over the next eight years, it raised over $445 million, counted Microsoft and the Qatar Investment Authority as investors, and crossed the $1.3 billion valuation mark.

By May 2025, it has gone bankrupt.

A wizard and a house of cards

Builder.ai was born in London, from co-founder Sachin Dev Duggal’s frustration with traditional software development. His solution: combine modular code components with human developers, coordinated by AI. The company set out to fix that by combining modular code libraries with human developers, guided by an AI layer. Its platform, dubbed “Builder Studio,” which came with a digital assistant “, Natasha”,  promised a seamless user experience powered by artificial intelligence.

What customers didn’t see: most of the work was being done by developers in India, not the AI. In 2019, The Wall Street Journal revealed the inconvenient truth: Builder.ai’s AI was more marketing gimmick than an engineering breakthrough. It was, as the report put it, “all engineer, no AI.” This gap between narrative and reality would define the company’s arc.

Built on fragile claims?

The AI pitch worked wonders with investors. By 2022, Builder.ai had raised $195 million, and in May 2023, it added another $250 million in a round led by QIA. The same year, Microsoft came on board as a strategic investor and partner, integrating Builder.ai’s platform into its cloud offerings. The validation was enormous. So were the expectations.

Internally, cracks were widening. The company had long relied on inflated revenue projections and AI claims to secure funding, according to insiders. A sprawling global workforce and costly expansion plans, including new markets in Southeast Asia and the Middle East, pushed the burn rate higher.

Then came the reckoning.

The domino effect

This month, Viola Credit, one of Builder.ai’s senior lenders, seized $37 million from the company’s accounts, triggering a default. CEO Manpreet Ratia, who had taken over just two months earlier to clean up the mess, was left with just $5 million in cash. A few days later, he filed for insolvency.

Turns out, Builder.ai had provided lenders with overstated financial projections, misrepresenting its revenue health. That breach of covenant gave Viola the green light to pull the plug. But the greater cause behind this structural collapse was that their business model never matched their branding. Ratia, in a company-wide call, acknowledged the writing on the wall. Most of its global staff were let go. Its product, once positioned as a flagship of AI innovation, was shelved.

The incident also reinforces concerns around “AI washing” — a practice where companies package conventional technology services as artificial intelligence to attract funding. Builder.ai, industry observers note, was a textbook example. Its failure has reignited conversations around the need for technical due diligence in AI deals.

For customers,  many of whom were startups and SMEs, the abrupt shutdown has created a scramble to rebuild or migrate their apps. It has underscored the risk of relying on emerging players for mission-critical software infrastructure.

Despite the blow, the broader low-code/no-code market remains resilient. Gartner projects that 60% of new enterprise apps will be developed using such platforms by 2028. The global market is expected to reach $26 billion by the end of this year.

From Gartner accolades to Fast Company rankings, from celebrity investors to top-tier logos on its website, Builder.ai appeared to be one of the AI age’s great success stories. But like many companies built on hype, it confused scale with sustainability, and visibility with viability. In the end, Builder.ai’s story is less about a technology that failed and more about the consequences of pretending it ever worked.