On 20 May 2025, Builder.ai’s CEO told employees on a hastily convened call that the company was insolvent. Creditor Viola Credit had seized $37 million from its accounts, triggering immediate collapse. Operations stopped overnight. Customer apps and data were lost. Three months later, Bloomberg confirmed what The Wall Street Journal had reported six years earlier: the AI was a fiction. Roughly 700 engineers in India had been writing the code while the company marketed an AI assistant called “Natasha.” Builder.ai had raised more than $450 million, including a strategic investment from Microsoft in May 2023 and a $250 million Series D led by the Qatar Investment Authority two weeks later. Peak valuation: $1.5 billion. The company is the largest AI-washing failure on record.

What The Postmortem Shows

Four diagnostic facts now in evidence.

Fact 1: The AI claim was challenged six years before collapse. The Wall Street Journal reported in August 2019 that Engineer.ai, as it was then known, used human engineers rather than AI for the bulk of its coding work. The story did not stop the next $400 million from arriving. Investors discounted the report as a competitive smear or a temporary state that the company would grow out of. It was neither. Former employees described the company in post-collapse interviews as “all engineer, no AI” — a state that persisted to the end.

Fact 2: The revenue was inflated by ~4x. Builder.ai claimed $220 million in revenue for 2024. Audits showed the real figure was closer to $55 million — a 75% overstatement. The gap came from booking projected future contracts as recognised revenue and from a round-tripping scheme.

Fact 3: There was an $180M round-tripping scheme with another Indian unicorn. Bloomberg reported in May 2025 that Builder.ai faked business with VerSe Innovation, the parent of DailyHunt, with the two companies invoicing each other for roughly the same amounts between 2021 and 2024 — in many cases for services that were never delivered. The aggregate was approximately $180 million of synthetic revenue across the period.

Fact 4: Indian engineers were instructed to behave like bots. The 700-person engineering team in India was directed to time their communications to UK business hours and avoid Indian-English colloquialisms so customers would not detect that humans, not software, were producing the deliverables. This was not aggressive marketing. It was operational deception.

Why It Survived

The deception ran for nine years against a backdrop of repeated public warning signs. Three forces kept it alive. First, the AI hype cycle made unverifiable AI claims fundable; investors signing $250M cheques in 2023 had no incentive to push hard on whether Natasha was real. Second, Microsoft’s strategic investment provided social proof that overrode the WSJ report — a $1.5B valuation supported by Microsoft is not a number that LPs question. Third, the round-tripping scheme made the revenue look like it was scaling on a curve that justified the valuation, when the underlying organic revenue was a fraction of the reported figure.

Why It Matters

Builder.ai is not an outlier event. It is the prototype for a category of failures that will surface across the 2023-2024 AI funding cohort. The diagnostic checklist that would have caught Builder.ai applies broadly: (1) is the “AI” replicable by humans at scale, and if so, what would the unit economics look like in that case? (2) does the customer cohort include companies whose only relationship is reciprocal invoicing? (3) is reported revenue tied to delivered services or to forward bookings? (4) are operational details (response times, communication patterns, output quality) consistent with the claimed automation?

The investors most exposed to AI-washing failures in the next eighteen months are the ones who funded their largest AI bets between mid-2023 and end-2024 without these checks running.

The Charaka View

Manthan Intelligence’s calibration database catalogues 220+ startup postmortems, and Builder.ai now joins the AI wrapper graveyard cohort — but with a sharper lesson. Wrapper companies died because their moat eroded; Builder.ai died because the moat was never there. Our analytical lenses now include an explicit “automation verifiability” score on any AI-claiming company assessment: if the claimed automation cannot be verified through technical due diligence, output sampling, or customer interview, the score caps the recommended exposure regardless of revenue trajectory. The 2019 WSJ report was the first signal. The 2024 revenue restatement was the last. Six years and four hundred million dollars passed between them. The cost of skipping diligence on AI claims is now visible. The next collapse is already in the data.

This analysis draws on Bloomberg’s reporting on Builder.ai’s bankruptcy, Bloomberg’s investigation of the VerSe round-tripping scheme, Rest of World’s coverage of the AI deception, and TechCrunch’s reporting on Microsoft’s strategic investment. Human editorial oversight applied.

This analysis is informational and does not constitute investment advice, a research report, or a recommendation to buy, sell, or hold any security.

Charaka Notes by Manthan Intelligence. Subscribe