The artificial intelligence boom has produced some of the fastest-growing startups in technology history, with companies routinely announcing tens or even hundreds of millions of dollars in annual recurring revenue (ARR) just a few years after launch. But behind many of those headline-grabbing figures, critics say a growing number of AI startups are relying on aggressive accounting practices that may significantly overstate their true business performance.
The debate erupted publicly after Scott Stevenson, co-founder and CEO of Spellbook, one of the most popular AI platform for contracts and transactional legal work, accused parts of the AI industry of engaging in what he described as a widespread effort to inflate revenue metrics in order to attract investment, secure media attention, and justify soaring valuations.
In a widely shared social media post, Stevenson argued that many AI companies reporting record-breaking growth are relying on revenue calculations that differ substantially from traditional definitions of ARR, while some investors and venture capital firms are allegedly aware of the practice and continue to promote the numbers publicly.
His comments reignited a long-running debate within Silicon Valley about how startup growth is measured and whether the AI boom is encouraging companies to stretch accepted financial metrics beyond their intended purpose.
The Metric at the Center of the Controversy
At the heart of the dispute is Annual Recurring Revenue, commonly known as ARR.
For decades, ARR has been one of the most closely watched performance indicators in the software industry. The metric emerged during the rise of cloud computing and subscription-based software services, providing investors with a simple way to estimate how much recurring revenue a company generates from active customers over a 12-month period.
Because many software companies collect payments gradually through monthly or annual subscriptions, ARR became a useful proxy for future revenue growth and business stability.
However, unlike audited financial statements governed by generally accepted accounting principles (GAAP), ARR is not formally regulated. As a result, companies have significant flexibility in how they calculate and present the metric. Critics argue that this flexibility is increasingly being exploited in the AI sector.
From ARR to “Committed” ARR
According to “tech crunch” report, many founders, investors, and startup finance professionals familiar with the industry, said that one of the most common practices involves replacing traditional ARR with a more speculative metric known as Contracted Annual Recurring Revenue (CARR), sometimes referred to as Committed ARR.
Contracted Annual Recurring Revenue (CARR) is a forward-looking financial metric used primarily by Software-as-a-Service (SaaS) and subscription-based companies. It represents the total annualized value of all subscription contracts that have been legally signed, regardless of whether the service has actually gone live or if the customer has started paying yet.
Under this approach, startups count revenue from signed customer agreements that have not yet been fully implemented or activated.
In theory, CARR can provide useful insight into future growth by highlighting revenue that is expected to materialize after deployment. In practice, however, critics say it introduces substantial uncertainty because many contracts never generate their full anticipated value.
Customers may delay implementation, reduce the scope of agreements, renegotiate terms, or cancel contracts entirely before significant revenue is collected.
Despite those risks, some companies reportedly present CARR figures publicly as standard ARR, creating the impression that the revenue is already active and recurring.
Industry insiders told TechCrunch that in certain cases, contracted revenue can exceed actual recurring revenue by as much as 70%, creating a substantial gap between public perception and underlying business performance.
The Pressure of the AI Boom
The controversy comes amid unprecedented competition among AI startups seeking funding from venture capital firms eager to identify the next industry giant. Unlike previous software cycles, where companies were expected to grow steadily over several years, AI startups are now under pressure to demonstrate explosive expansion almost immediately after launch.
The rise of generative AI has fueled expectations that startups can achieve revenue milestones once considered extraordinary within remarkably short timeframes. As investors search for category leaders, founders face growing pressure to produce increasingly impressive growth metrics.
Industry observers say this environment has created powerful incentives to present the strongest possible revenue narrative. “The valuations have gotten higher, and so the incentives are stronger,” according to venture capital executives familiar with the sector.
In many cases, revenue announcements have become central to determining which startups attract media attention, recruit top talent, secure customers, and command billion-dollar valuations.
When Future Revenue Becomes Present Revenue
Some of the examples described by industry sources illustrate how blurred the distinction between actual and projected revenue has become.
Several investors told TechCrunch they were aware of high-profile enterprise AI companies that publicly claimed to have surpassed $100 million in ARR despite generating only a portion of that amount from active paying customers.
The remainder reportedly consisted of signed contracts that had not yet been fully implemented and, in some cases, might take months or years before producing significant revenue.
Former employees at some startups also described situations where companies counted lengthy free pilot programs as ARR by including future payments expected after the trial period ended.
The problem, critics argue, is that these future payments remain uncertain until customers complete implementation and begin paying. As a result, investors and observers may receive an overly optimistic picture of a company’s actual financial position.
Another Source of Confusion: Annualized Run-Rate Revenue
Complicating matters further is the existence of another metric that shares the same ARR acronym: Annualized Run-Rate Revenue.
Annualized Run-Rate Revenue (ARRR) is a financial metric used to estimate how much revenue a company would generate over a full year if its current revenue pace continues unchanged.
It is especially common among startups, SaaS companies, and AI firms because it allows them to showcase growth before they have a full year of operating history. If a company generated a certain amount of revenue during a recent period (a month or a quarter), it “annualizes” that figure by multiplying it to project a full year’s revenue.
Unlike traditional ARR, which is based on recurring contracts, annualized run-rate revenue extrapolates recent performance into a future annual figure.
For example, a startup generating $1 million in revenue during a single month could theoretically claim a $12 million annualized run rate, assuming growth remains constant.
The approach can be particularly problematic in AI, where many companies charge customers based on usage rather than fixed subscriptions.
Revenue generated during periods of unusually high demand may not be sustainable over an entire year, making annualized projections vulnerable to significant overstatement.
Critics argue that mixing multiple definitions of ARR further muddies the waters for investors, journalists, and prospective customers attempting to assess a company’s true performance.
Venture Capital’s Role Under the Microscope
The controversy has also placed venture capital firms under increased scrutiny.
Several founders and industry executives argue that investors often have little incentive to challenge aggressive revenue reporting because inflated growth metrics can benefit their portfolios.
When startups announce rapid revenue growth, they are more likely to attract additional funding, generate positive media coverage, and establish themselves as leaders within their sectors.
That momentum can ultimately boost the valuation of both the startup and the venture firms backing it.
According to critics, some investors may therefore tolerate aggressive interpretations of ARR even when they privately understand the limitations of the figures being reported.
Industry participants suggest that venture capital firms frequently have access to more detailed financial information than the public, allowing them to distinguish between actual recurring revenue and more speculative measures while still benefiting from the publicity generated by larger headline numbers.
Growing Skepticism Across the Industry
Not everyone in the startup ecosystem is comfortable with the trend. Some founders argue that exaggerated revenue reporting risks creating unrealistic expectations and could ultimately damage confidence in the broader AI sector.
Executives who lived through previous technology cycles point to the lessons of past market corrections, when companies that relied on aggressive growth narratives struggled to justify their valuations once investors demanded stronger financial fundamentals.
For these founders, transparency is becoming a competitive advantage rather than a liability. They argue that while inflated metrics may deliver short-term publicity and higher valuations, they also increase the likelihood of future disappointment when companies are eventually judged by traditional financial measures.
A Potential Reckoning Ahead
The dispute over ARR reflects a broader tension emerging across the AI economy.
On one side are founders and investors racing to capitalize on one of the largest technology booms in decades. On the other are growing concerns that competition for capital and attention is encouraging increasingly aggressive interpretations of business performance.
As AI startups continue raising billions of dollars and pursuing record valuations, scrutiny of financial metrics is likely to intensify.
For now, ARR remains one of the most powerful numbers in Silicon Valley. But as critics warn, the question investors may increasingly need to ask is not how much ARR a company reports—but what, exactly, that ARR actually represents.