source: techcrunch ai: how vcs and founders use inflated ‘arr’ to crown ai startups
level: business
scott stevenson, ceo of legal ai startup spellbook, recently called out a widespread practice among ai startups: inflating annual recurring revenue figures. he described it as a huge scam, noting that many record-breaking revenue claims rely on a dishonest metric. the core issue is substituting contracted arr, or carr, for actual arr. carr includes revenue from signed customers who are not yet onboarded, making it a squishier number that can be far higher than real recurring revenue.
techcrunch spoke with over a dozen founders, investors, and finance professionals who confirmed that fudged arr is common. one vc said he has seen companies where carr is 70% higher than arr, with much of that contracted revenue never materializing. some startups count lengthy free pilots or heavily discounted early years as full arr. investors often know about the exaggerations but stay silent because high public revenue numbers help portfolio companies attract talent and customers, creating a narrative of runaway winners.
the pressure to show explosive growth in ai has intensified these practices. some founders use annualized run-rate revenue, extrapolating from a short period, which can be misleading for usage-based models. not all startups engage in this; some prioritize transparency, warning that inflated metrics create higher hurdles and can backfire. the phenomenon is not new, but ai hype and soaring valuations have made it more aggressive, with vcs incentivized to overlook or support the overstatements.
why it matters: inflated revenue metrics distort market perceptions, making it harder for data scientists and ai professionals to assess a startup's true health and scalability.
source: techcrunch ai: how vcs and founders use inflated ‘arr’ to crown ai startups