How AI startups are using CARR to mask actual revenue

Last month, Scott Stevenson, the co-founder and CEO of the legal AI startup Spellbook, sparked an industry-wide debate by labeling the public revenue inflation among AI startups a “huge scam.” Stevenson took to X to criticize what he described as “ARR shenanigans,” where companies sometimes substitute Contracted Annual Recurring Revenue (CARR)—which includes committed but not yet active contracts—for traditional Annual Recurring Revenue (ARR). Stevenson wrote: “The reason many AI startups are crushing revenue records is because they are using a dishonest metric. The biggest funds in the world are supporting this and misleading journalists for PR coverage.” His post drew over 200 reshares and comments, highlighting growing concern over how software metrics are reported.

The core of the issue lies in how startups account for future revenue. While ARR historically tracks the annual value of active, paying customer contracts, CARR is a squishier metric that counts revenue from signed contracts before the product is actually implemented or deployed. Bessemer Venture Partners explained in 2021 that CARR builds on ARR by adding committed but not yet live contract values. In many cases, investors are aware of these exaggerations in public revenue declarations. Some venture capitalists report seeing startups with CARR figures 70% higher than actual ARR. This practice allows companies to publicly claim they have surpassed major milestones, such as $100 million in ARR, even when only a fraction of that revenue comes from currently paying customers. In other instances, the discrepancies are smaller but still notable; one startup reportedly claimed $50 million in ARR in its marketing materials when its actual ARR was $42 million, leaving an $8 million gap.

Venture capitalists are often incentivized to support these public declarations to build a narrative of “runaway winners” and secure press coverage. Jack Newton, the co-founder and CEO of legal startup Clio—which was valued at $5 billion last fall—noted that Stevenson did a great job of highlighting what some might describe as “bad behavior” by these companies. Newton added that some investors look the other way when their portfolio companies inflate numbers because it improves outward perceptions. However, this practice creates an environment that feels “fake” to industry insiders, according to Alex Cohen, the co-founder and CEO of health AI startup Hello Patient. While traditional accounting standards like GAAP (Generally Accepted Accounting Principles) focus on historical, collected revenue rather than future projections, the lack of formal audits for ARR allows these public misrepresentations to persist.

Why it matters

AI startups are increasingly inflating their revenue metrics by using “contracted ARR” (CARR) instead of traditional ARR, a practice that investors are often aware of but overlook to maintain narratives of rapid growth.