How VCs and founders use inflated ‘ARR’ to crown AI startups

In a recent revelation, Scott Stevenson, the co-founder and CEO of the legal AI company Spellbook, publicly criticized what he describes as rampant deception in the AI startup sector regarding revenue reporting. Taking to social media, he labeled the inflation of revenue figures as a significant scam, suggesting that many AI startups are achieving record revenues by utilizing misleading metrics. Stevenson’s claims have sparked a lively debate within the community, attracting attention from prominent investors and fellow founders. Stevenson argues that the manipulation of Annual Recurring Revenue (ARR) is a widespread issue, pointing out that some startups are confusing contracted ARR, also known as Committed ARR (CARR), with true ARR. This confusion allows companies to report inflated revenue figures that do not accurately reflect their actual earnings. One investor noted that when one startup adopts this practice, others in the same sector feel pressured to follow suit to remain competitive. ARR has long been a trusted metric in the tech industry, particularly during the cloud computing boom. It is designed to reflect the total revenue from all active customer contracts. However, since it is not formally audited, the accuracy of ARR can be questionable. The term typically refers to revenue that has been collected, while CARR includes contracts that are signed but not yet fulfilled, making the latter a less reliable indicator of a startup’s financial health. Some investors have reported seeing discrepancies where a startup claims to have exceeded $100 million in ARR, yet only a fraction of that amount was actually coming from paying customers. The rest was composed of contracts that were either not yet implemented or involved lengthy trial periods. This practice raises significant concerns, as it can lead to inflated valuations and mislead potential investors and customers. Critics like Ross McNairn, CEO of the legal AI firm Wordsmith, echo Stevenson’s sentiments, stating that there are concerning trends in how startups present their revenue figures. He pointed out that while some firms are transparent about their metrics, others engage in questionable practices that could ultimately damage their credibility. The issue is further complicated by the emergence of another revenue metric — annualized run-rate revenue — which can also lead to misinterpretation. This method extrapolates current revenue over the next year, but in a landscape where usage-based fees are common, it can create an unrealistic picture of future income. As the pressure mounts for startups to demonstrate rapid growth in the AI sector, some venture capitalists may overlook discrepancies in revenue reporting to maintain a favorable narrative. This complicity can perpetuate a cycle where inflated figures become the norm, creating a bubble that could eventually burst. While some startups strive for transparency and resist the temptation to inflate their numbers, the allure of attracting top talent and investment through exaggerated claims is powerful. As the conversation around revenue integrity continues, many insiders express skepticism over the claims of rapid growth, questioning the sustainability of such practices in the long run.