"A Huge Scam": Why Tech Insiders Say AI Startup Revenue is Wildly Inflated

The Hidden Math Behind the AI Boom: Inside the Industry’s Secret Revenue Game:

The ARR Illusion: How AI Startups and Their Investors Are Inflating Revenue Metrics Inside the widespread practice of ARR manipulation, committed ARR inflation, and the metrics game reshaping AI startup valuations.

The Revenue Reality Check: What AI Startups Are Really Reporting:

A bombshell accusation ignited the AI startup world when Scott Stevenson, co-founder and CEO of the legal AI startup Spellbook, publicly called out what he described as a "huge scam" sweeping through the AI industry. In a viral post on X, Stevenson accused high-profile AI startups — backed by some of the biggest venture capital funds in the world — of systematically inflating their revenue figures to manufacture headlines and attract investors.

"The reason many AI startups are crushing revenue records is because they are using a dishonest metric," Stevenson wrote, adding that "the biggest funds in the world are supporting this and misleading journalists for PR coverage." The post struck a nerve, drawing over 200 reshares and sparking responses from investors, founders, and journalists across the startup ecosystem.

Understanding ARR: The Metric That Built the SaaS Era:

Annual Recurring Revenue (ARR) has long been the gold standard metric for measuring the financial health of subscription-based and SaaS (Software as a Service) companies. Established during the cloud computing era, ARR was designed to reflect the total annualized value of active, paying customers under contract — a reliable snapshot of predictable revenue.

The beauty of traditional ARR lies in its simplicity and honesty. It counts only signed-and-sealed sales — money from customers who are live, onboarded, and actively paying. Accountants don't formally audit ARR because it falls outside Generally Accepted Accounting Principles (GAAP), which focus on already-collected historical revenue. But its informal trustworthiness made it the benchmark metric for startup growth and investor confidence.

Fast forward to the AI startup boom, and that trustworthiness is being aggressively exploited. As AI companies race to announce eye-popping ARR milestones — $50M, $100M, and beyond — many are quietly redefining what those numbers actually mean, substituting a far squishier metric in its place.

The CARR Problem: Committed ARR vs. Actual Revenue:

The most common manipulation tactic, according to multiple sources, is substituting "Committed ARR" (CARR) — also called "Contracted ARR" — for traditional ARR, while publicly labeling the figure simply as "ARR." This subtle sleight of hand allows startups to count revenue from signed contracts where the customer hasn't even been onboarded yet.

"For sure they are reporting CARR as ARR. When one startup does it in a category, it is hard not to do it yourself just to keep up." — Anonymous Investor

CARR was introduced as a legitimate forward-looking growth tracker. According to Bessemer Venture Partners (BVP), which defined the metric in a 2021 blog post, CARR "builds on the ARR concept by adding committed but not yet live contract values to total ARR." Critically, BVP noted that CARR should be adjusted to account for expected customer churn and downsell — customers who cancel or reduce their spending.

The fatal flaw emerges when CARR is reported without those critical adjustments. One VC told TechCrunch that they had personally seen companies where CARR was 70% higher than actual ARR — and a significant chunk of that contracted revenue never materialized. Clients can cancel contracts during lengthy implementation phases, before they've paid a single dollar.

The most extreme cases paint a deeply troubling picture. Multiple investors told TechCrunch they directly know of at least one high-profile enterprise AI startup that publicly announced it had surpassed $100 million in ARR, when only a fraction of that figure came from currently paying customers. The remainder was from contracts that hadn't been deployed yet — and in some cases faced implementation timelines stretching months or years.

Inside the Scam: Free Pilots Counted as Paid Revenue:

Perhaps the most egregious example uncovered in TechCrunch's reporting came from a former employee at a startup that routinely reported CARR as ARR. According to this source, the company counted at least one substantial, year-long free pilot program as ARR — revenue that hadn't been collected, and might never be.

What makes this particularly alarming is the boardroom awareness. The company's board of directors — including a venture capitalist from a major fund — was reportedly aware that the eventual paying portion of the contract had been included in ARR figures during the lengthy, still-unpaid pilot period. The board also knew the customer retained the right to cancel before the full contract amount was paid.

"I think Scott is right. I've heard all sorts of anecdotes as well. I speak to VCs all the time. They're like, 'There are some choppy, choppy standards out there.'" — Ross McNairn, Co-founder & CEO, Wordsmith

Not every case reaches this level of severity, but the pattern is pervasive. One employee at a different startup described a more modest but still misleading gap: the company's marketing materials claimed $50 million in ARR, while the actual figure was $42 million. Investors with access to the company's books knew the real number — but were reportedly comfortable with the discrepancy, viewing an $8 million gap as a "rounding error" the company would quickly grow past.

The Second ARR Problem: Annualized Run-Rate Revenue:

Committed ARR isn't the only ARR manipulation tactic in play. There's a second, entirely different metric that also goes by the "ARR" acronym: Annualized Run-Rate Revenue. This calculation extrapolates current revenue over the next 12 months based on a recent period — a quarter, a month, a week, or even a single day of strong performance.

For AI companies that charge based on usage or outcomes, this method is particularly susceptible to inflation. Unlike traditional subscription revenue locked into multi-year contracts, usage-based revenue can spike dramatically during a hot week and collapse just as quickly. Reporting a week of peak usage as annualized ARR is not just misleading — it's a distortion of financial reality that can lead investors and the media to dramatically overestimate a company's sustainable revenue base.

The confluence of these two ARR distortions — CARR misrepresentation and run-rate inflation — creates a perfect storm for startup valuation manipulation. When both tactics are deployed simultaneously, a startup's publicly reported "ARR" could be multiples above its actual recurring revenue from paying customers.

Why Investors Tolerate — and Enable — ARR Inflation:

The uncomfortable truth, according to sources interviewed for this story, is that ARR inflation often serves the interests of investors just as much as founders. Inflated metrics generate press coverage, elevate a startup's perceived market position, and can help justify higher valuations in subsequent funding rounds — all of which benefit existing investors.

The competitive dynamics of AI sectors make the problem self-reinforcing. As one investor explained, when one startup in a category begins reporting CARR as ARR, competitors feel pressure to do the same simply to maintain perceived parity. Refusing to play the game can mean appearing to fall behind in a race that's partly imaginary.

Jack Newton, co-founder and CEO of legal tech startup Clio, told TechCrunch that Stevenson's viral post brought "much-needed awareness" to the topic. Newton pointed to a post by Y Combinator's Garry Tan as an example of industry leaders calling for clearer, more honest revenue metric standards. The fact that such calls are necessary underscores how widespread the problem has become.

The GAAP Gap: Why ARR Manipulation Goes Unchecked:

A structural vulnerability in startup accounting makes this kind of manipulation easy to sustain. Because ARR is not a GAAP-defined metric, it is never formally audited or independently verified in the way that revenue on a public company's income statement would be. Startups are free to define, calculate, and report ARR however they choose — and many are choosing to stretch the definition as far as it will go.

The concept most similar to CARR under formal accounting standards is "remaining performance obligations" — a GAAP measure that tracks the value of contracts yet to be fulfilled. But remaining performance obligations come with rigorous disclosure requirements and churn adjustments that CARR, as commonly used by AI startups, conspicuously lacks. Without mandatory standards, the ARR arms race will likely continue. Most experts interviewed for this story noted that ARR manipulation is not a new phenomenon — but AI-era hype has turbocharged the aggressiveness with which startups and their investors are willing to stretch the metric.

What Honest ARR Reporting Should Look Like:

Reform advocates in the startup community argue that transparency starts with labeling. If a startup is reporting CARR rather than traditional ARR, it should say so explicitly — and it should disclose the churn and downsell assumptions baked into that figure. Similarly, annualized run-rate revenue should always be clearly labeled as a run-rate estimate, not reported as if it were locked-in contracted revenue.

The Bessemer Venture Partners framework provides a useful template. BVP's original definition of CARR explicitly requires adjustment for expected churn and downsell, transforming it from a vanity metric into a genuinely useful forward indicator. Adopting that standard broadly — and enforcing it through investor due diligence — would go a long way toward restoring trust in startup revenue figures.

Ultimately, the responsibility lies with investors and founders alike. When boards of directors knowingly allow inflated metrics to be published externally while tracking accurate numbers internally, they are prioritizing short-term PR wins over the long-term credibility of their companies. As AI startup valuations face increasing scrutiny, the startups that establish a reputation for rigorous, honest reporting may find that integrity becomes a competitive advantage.

The Bottom Line: ARR Inflation Is a Systemic Problem:

The ARR manipulation scandal illuminates a deeper issue in the AI startup ecosystem: in a market driven by hype, narrative, and competitive optics, the pressure to report impressive numbers can overwhelm the incentive to report accurate ones. When the biggest funds in the world are reportedly complicit in misleading metrics, the problem is not a few bad actors — it's a systemic failure of accountability.

For journalists, potential customers, and future investors evaluating AI startups, the lesson is clear: always ask what's behind the ARR number. Is it actual recurring revenue from paying customers? Is it contracted revenue from clients not yet onboarded? Is it a run-rate extrapolation from a single week of peak activity? The answer could mean the difference between a thriving business and an expensive illusion.

As Scott Stevenson's viral post demonstrated, the AI startup community is capable of self-correction — but only when someone is willing to name the problem publicly.

In a sector moving at extraordinary speed, the courage to demand honest metrics may be one of the most important forms of accountability available.

Topics: ARR Inflation • AI Startups • Venture Capital • Startup Metrics • CARR • SaaS Revenue • Startup Accounting • Annualized Run Rate • AI Funding