Last month, Scott Stevenson, founder and CEO of AI startup Spellbook, took on X in an attempt to expose what he called the “biggest scam” among AI startups: the inflation of the financial figures they publicly declare.
“The reason many AI startups are breaking revenue records is because they are using an unreliable metric. The biggest funds in the world support this and mislead PR journalists,” he wrote in his tweet.
Stevenson is not the first to point out that annual recurring revenue (ARR) – a metric historically used to sum up the annual revenue of active customers under contract – is being used by AI companies in an unrecognizable way. Certain aspects of ARR shenanigans have been the subject of many news reports and social media posts.
However, Stevenson’s tweet seems to have struck a chord within the AI startup community, drawing more than 200 reshares and comments from top investors, several inventors, and several articles.
“Scott at Spellbook has done a great job of highlighting what you might describe as bad behavior in other companies,” Jack Newton, founder and CEO of legal firm Clio, told TechCrunch, adding that the post brought much-needed awareness to the topic, referring to a post with an explanation from YC’s Garry Tan about relevant revenue metrics.
TechCrunch spoke to more than a dozen founders, investors, and startup finance experts to assess whether ARR inflation is as pervasive as Stevenson suggests.
Indeed, our sources, most of whom spoke on the condition of anonymity, confirmed that the ARR included in public announcements is common among startups, and how, in many cases, investors are aware of the exaggeration.
It’s not really income, yet
The main obfuscation trick is to replace the “contracted ARR,” sometimes called “contracted ARR” (CARR), and simply call it ARR.
“They definitely report CARR” as ARR, said one investor. “When one startup makes it to the division, it’s hard not to do it yourself to keep going.”
ARR is an established and trusted metric from the cloud era to show the total sales of products where usage, and payments, are calculated over time. Accountants do not formally review or sign off on ARR primarily because generally accepted accounting principles (GAAP) focus on historical revenue, which has already accrued, rather than future revenue.
ARR was intended to reflect the total number of signed and closed sales, usually multi-year contracts. (Today, this concept often goes by another name: residual operating obligations.) Currently, the term “revenue” is generally reserved for accrued income.
CARR should be another way to track growth. But it’s a squishier metric than ARR because it counts revenue from signed-up customers who haven’t logged in yet.
One VC told TechCrunch that he has seen companies where the CARR is 70% higher than the ARR, even though most of that contract money will never be seen.
CARR “builds on the ARR concept by adding committed but not live contract values to the overall ARR,” Bessemer Venture Partners (BVP) wrote in a blog post back in 2021. Critically, though, BVP says, startups should adjust CARR to take into account expected customer churn (how many customers leave) and “downsell” (those who decide to buy less).
The biggest problem with CARR is calculating the revenue before the startup product is used. If usage takes too long or goes poorly, customers may cancel during the trial period before all – or some – of the contract revenue is collected.
Several investors told TechCrunch that they are directly aware of at least one high-profile startup that reported surpassing $100 million in ARR, while only a fraction of that revenue came from current paying customers. Some were from contracts that had not yet been entered into and in some cases it would take a long time to implement the technology.
Another former employee who used to report CARR as ARR told TechCrunch that the company counted at least one major, year-round pilot as ARR. The company’s board, including the VC of the big fund, knew that the proceeds from the last paid portion of the contract were counted in ARR during a lengthy evaluation process, the person said. The board also knew that the customer could cancel before paying the full amount of the contract.
The obvious problem with using CARR and calling it ARR is that it is more susceptible to being “played” than regular ARR. If the startup doesn’t realistically account for churn and downsell, the CARR can go up. For example, a startup might offer big discounts for the first two years of a three-year contract and calculate all three years as CARR (or ARR), even though customers might not stick around to pay higher rates in the third year.
“I think it’s Scott [Stevenson] it’s okay. I’ve also heard all kinds of anecdotes,” Ross McNairn, founder and CEO of legal AI startup Wordsmith told TechCrunch about ARR’s misrepresentations. “I talk to VCs all the time. It’s like, ‘There are strange, strange values.’
Most cases are extremely minor. For example, an employee at one startup explained the difference when marketing materials cost $50 million in ARR, while the actual cost was $42 million.
However, the person said that investors have access to the company’s books, which accurately reflect the lower value. The source said some startups and their investors are free to play fast and loose with their public metrics in part because AI startups are growing so fast that an $8 million gap is considered a rounding error and they will grow faster.
Another, “ARR” is more problematic
There is another issue surrounding all those public ARR announcements. Sometimes founders use another measure with the same name as “ARR” and the same name: annual running average revenue.
This ARR is also controversial because it extrapolates current revenue over the next 12 months based on revenue for a given time period (eg, quarter, month, week, or day).
Since many AI companies charge based on usage or results, that method of calculating annual performance ARR can be misleading because revenue is no longer tied to predictable contracts.
Most of the people interviewed for this story said that ARR exaggeration of all kinds is nothing new, but startups have become more aggressive amid the AI hype.
“The numbers are so high, so the incentives are strong to do it,” Michael Marks, founding managing partner at Celesta Capital, told TechCrunch.
In the era of AI, startups are expected to grow faster than ever.
“Going from 1 to 3 to 9 to 27 is not interesting,” said Hemant Taneja, CEO and managing director of General Catalyst, on the 20VC podcast last September, referring to the billions of ARR the typical startup is expected to hit each year. “You have to go like 1 to 20 to 100.”
The pressure to show rapid growth causes some VCs to support, or at least ignore, startups that present inflated ARR figures to the public.
“Certainly there are VCs in this because they are motivated to make a story that they have run away winners. They are motivated to get the stories of their companies,” Stevenson told TechCrunch.
Newton, whose first official AI Clio was valued at $5 billion last fall, also says VCs are often aware but silent about ARR’s negative distortions. “We see some investors looking the other way when their companies are raising numbers because it makes them look good on the outside,” he told TechCrunch.
What VCs really think
Some investors who spoke to TechCrunch said there is no reason for VCs to make exaggerated statements.
By ignoring public announcements of increased ARR, VCs are effectively helping to make their portfolio companies. When a startup publicly reports high revenue, it is more likely to attract the best talent and customers who believe the company is the undisputed winner in its category.
“Investors can’t afford it,” the VC told TechCrunch. “Everybody has a company that makes money with CARR like ARR.”
Still, anyone familiar with the intricacies of the industry finds it hard to believe that some of these startups have actually reached $100 million in ARR within a few years of launch.
“For everyone on the inside, it feels like a lie,” said Alex Cohen, founder and CEO of health AI startup Hello Patient. “You read the headlines and say, ‘I don’t believe it.'”
However, not all startups feel comfortable representing growth by reporting CARR instead of ARR. They prefer to be clean and clear with their numbers in part because they understand that the public markets value software companies on ARR rather than CARR. These founders prioritize transparency.
Wordsmith’s McNairn, who remembers the initial struggles he faced due to the eligibility of higher figures after the 2022 market adjustment, said he did not want to create an even bigger obstacle by exaggerating his initial profit.
“I think it’s short-sighted, and I think if you’re doing things like that for short-term gain, you’re overhyping multiples that are already crazy,” he said. “I think it’s bad hygiene, and it’s going to come back to bite you.”
If you shop through links in our articles, we may earn a small commission. This does not affect our editorial independence.
