AI firms face scrutiny as Annual Recurring Revenue becomes Silicon Valley's least trusted metric

Silicon Valley Metric Under Fire

A controversy erupted when Roy Lee, CEO of Andreessen Horowitz-backed Cluely, admitted to fabricating his company's Annual Recurring Revenue figures. Lee initially told a TechCrunch reporter that Cluely's ARR doubled in one week to $7 million, only to later retract on X, calling it "the only blatantly dishonest thing I've said publicly online" and revealing the actual number was $5.2 million 1

1

. While Lee's boldness stands out, Silicon Valley investors say taking liberties with ARR has become common practice among AI firms, transforming what was once a reliable benchmark into one of the least trusted yardsticks for gauging startup growth.

The Wild West of Startup Metrics

"The startup world has always been a bit more of a Wild West," said Chuck Eesley, a professor of management science and engineering at Stanford University. "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. So essentially, the number can mean whatever the founder needs it to mean when they walk in to do a deal or do a fundraise" 1

1

. This lack of oversight has created an environment where scrutiny of ARR metrics has intensified, particularly as the measurement becomes ubiquitous across AI startup business models. Companies including OpenAI, Anthropic, Glean, and Cursor have all reported ARR for individual products or overall sales, making the metric a closely watched number for media outlets and investors alike 1

1

.

How ARR Calculations Work and Why They're Problematic

The basic ARR calculations are straightforward: take one month's revenue from recurring contracts and multiply by 12 for an annual projection 2

2

. Until recently, ARR was considered a trusty benchmark for software businesses, particularly those selling predictable services to other businesses. "This worked really well when subscription pricing was very straightforward," said Darren Yee, a senior venture associate at NYU's Innovation Venture Fund. "And that's been true for a long time, basically up until AI" 1

1

. The untrustworthiness of Annual Recurring Revenue stems from the leeway in how exactly to measure it—what contracts count, what time period to use—making it relatively easy for startups to massage figures, especially when revenue fluctuates week to week or subscription models lapse 1

1

.

The AI-Specific Challenge: Usage-Based Pricing

The shift toward usage-based pricing has made ARR particularly unreliable for AI firms. "Customers may have a nominal subscription number but are paying mostly for usage. This gives very lumpy revenue attribution in the early days," Yee explained. "You can't just take one month of subscriptions and multiply by 12 and get what that represents in an annual contract, because it probably won't play out that way" 1

1

. Additionally, many AI business customers eagerly try new tools during trial periods but drop them afterward. This trial revenue can be counted as "recurring," even though contracts don't renew, contributing to inflating ARR figures 2

2

.

What This Means for Fundraising and Due Diligence

The unreliability of ARR metrics places additional pressure on VCs to conduct thorough due diligence for VCs during fundraising rounds. Lee himself questioned the metric's validity for young companies, stating his measurement was changing 20% week to week 1

1

. Yet despite these concerns, more elaborate auditing processes could burden small startups. "I think we should be careful about imposing a lot of auditing and accounting costs on small startups and stifling a lot of the innovation and experimentation that should be going on," Eesley cautioned 1

1

. Investors and acquirers must now navigate this tension, relying on deeper analysis rather than surface-level metrics to assess true startup growth potential in an era where the traditional measures no longer capture the complexity of AI business models.