Many artificial intelligence (AI) startups use annual recurring revenue as a metric for their growth.
However, this increasingly common practice has become one of the least-trusted measurements of success in the AI space, Bloomberg News reported Tuesday (April 7).
“The startup world has always been a bit more of a Wild West,” Chuck Eesley, professor of management science and engineering at Stanford University, told Bloomberg.
“There are no audit requirements, there are no SEC definitions, so basically there’s no cop on the beat other than the VCs and acquirers doing their due diligence,” he said, so that a number can “mean whatever the founder needs it to mean” during dealmaking or fundraising.
Annual recurring revenue (ARR) calculations work like this: a company takes a single month’s revenue from recurring contracts and multiples it by 12 for a year-long projection. Companies that have used this practice include Anthropic and OpenAI.
The Bloomberg report noted that there is nothing inherently wrong with measuring growth this way, and if a company adds new subscribers each month, it can offer a more accurate glimpse into revenue than looking at past sales.
Darren Yee, a senior venture associate at NYU’s Innovation Venture Fund, told Bloomberg that—until recently—ARR was seen as a reliable gauge for software businesses, especially those selling predictable services to other companies.
“This worked really well when subscription pricing was very straightforward,” Yee said. “And that’s been true for a long time, basically up until AI.”
However, recurring revenue gives companies a lot of freedom in how precisely to measure, meaning it’s easy for startups to juke their numbers, which can vary if revenue swings from week to week or recurring subscriptions lapse.
Eesley said many AI business customers are eager to try new tools, but only on a trial basis before cancelling their subscriptions. Revenue from that trial period can be considered as “recurring,” even if a contract doesn’t renew.
In other AI news, OpenAI issued a report this week calling for an industrial policy that can manage the challenges presented by the technology.
“As AI reshapes work, knowledge, and production, incremental updates won’t be enough,” the company said in a LinkedIn post announcing the report. “The scale of change demands new ideas and institutions to ensure this transition benefits everyone.”