Inflated ARR at AI startups: how to spot it

The context: why ARR became the signature metric of AI

In the funding cycle of tech startups, ARR — Annual Recurring Revenue — has long been the main indicator of financial health. In the AI segment, however, this metric has taken on outsized weight. Investors use it to compare companies at different stages. Founders present it as proof of traction. The media pick it up to build growth narratives.

As a result, ARR has become a positioning tool before it is a measurement tool. At AI startups in particular, where sales cycles are still immature and contracts are often experimental, the pressure to show big numbers is very high. Consequently, some players have started interpreting the definition of ARR creatively.

According to a TechCrunch investigation from May 2026, founders and venture capitalists are fully aware of this practice. So these are not accidental accounting errors, but deliberate narrative choices.

The numbers that matter: how ARR gets inflated

There are several techniques an AI startup can use to report an ARR higher than the real one. First of all, it helps to understand them in detail so you can recognize them.

• Annualizing pilot contracts: a three-year, €30,000 contract is reported as €10,000 of ARR, even though renewal isn't guaranteed.
• Including cloud credits: free credits offered by AWS, Google Cloud or Azure are counted as actual revenue.
• Reclassified one-off revenue: onboarding payments or initial consulting fees are spread over a year and added to ARR.
• Conditional contracts: agreements subject to milestones or exit clauses are included in the total as if they were confirmed.
• Pipeline as ARR: in some cases, late-stage sales opportunities are added to actual figures.

In addition, some startups distinguish between committed ARR and run-rate ARR without making the difference explicit in public communications. Likewise, the term “ARR” is used interchangeably with “annualized revenue,” which is a different concept.

According to Gartner research on valuation metrics for AI startups, the lack of shared standards in defining ARR is one of the main sources of opacity in the sector. Moreover, no body certifies the consistency of these numbers before they are published.

Strategic view: why investors let it slide

One notable point from the investigation is that venture capitalists are not unwitting victims of this mechanism. On the contrary, they often encourage it or actively tolerate it. The reason is structural.

A fund that has invested in an AI startup wants that startup to be seen as a market leader. So a high ARR — even one built on shaky foundations — helps attract new investors in later rounds, raises the valuation and makes for more profitable exits. As a result, the incentive to correct the narrative is weak.

Moreover, in the 2026 AI market, perceived growth speed often matters more than the strength of the fundamentals. This creates an environment where companies reporting more conservative numbers risk being penalized in fundraising, even if their financial position is stronger. It's a systemic problem, not an individual one.

An analysis by Harvard Business Review on AI startup metrics had already shown last year how competitive pressure in funding rounds is distorting the quality of information available to outside stakeholders.

Operational implications for Italian B2B SMEs

For an Italian SME operating in B2B or retail, inflated ARR has direct, practical consequences. Many companies end up evaluating AI startups as potential technology suppliers, business partners or emerging competitors. In all these cases, the reported ARR shapes the perception of reliability and strength.

However, choosing an AI vendor based on inflated ARR means relying on a company whose customer base may be far more fragile than it looks. The risks are many: service disruption, renegotiation of contract terms, weak post-sale support.

At SHM Studio, we see that this topic still gets too little attention among Italian SMEs. Technology purchasing decisions are often made by looking at brand awareness or the numbers reported in the media, without a critical analysis of the underlying metrics. Building an in-house ability to critically read AI startups' financial data is now a strategic skill.

In particular, we suggest considering the following operational criteria when evaluating an AI partner:

• Ask for an explicit definition of ARR: a solid vendor can explain exactly what is included in its reported ARR.
• Check the number of active customers: a high ARR from a few customers is far riskier than one spread across a broad base.
• Analyze the churn rate: the percentage of customers who don't renew is a more reliable indicator of business health than raw ARR.
• Request verifiable references: documented use cases with companies of comparable size and industry are more informative than any aggregate metric.
• Assess the contract structure: prepaid annual contracts signal a different level of stability than monthly or conditional ones.

The work still ahead: toward more transparent standards

The inflated ARR issue isn't going to resolve itself in the short term. As long as capital markets reward perceived growth over real strength, the incentive to report optimistic numbers will remain structural. Still, some signs suggest the landscape is changing.

On one hand, institutional investors and private equity funds are starting to demand more granular revenue metrics during due diligence. On the other, some standardization initiatives — such as those promoted by SaaS industry organizations — are trying to define shared guidelines for ARR reporting.

Moreover, with possible European regulations on the transparency of startups' financial communications on the horizon, pressure for stricter standards is likely to grow by 2027-2028. Even so, in the short term, critical reading skills remain the only tool available to those operating outside the circle of institutional investors.

For a deeper look at how SMBs can critically evaluate AI tools, it's also worth reading Wired's analysis of the metrics problems in AI companies.

How SHM Studio supports SMEs in evaluating technology

At SHM Studio, we help Italian SMEs build a solid digital strategy, including a critical selection of tools and technology partners. That's why evaluating AI startups falls squarely within our AI consulting services.

When we support a client in choosing AI-based solutions, we go beyond features. We look at the vendor's stability, the contract structure, the product roadmap, and how well market promises match operational reality. This approach ties in with our digital marketing and SEO work, where the quality of the tools you adopt directly affects results.

If you're building or improving your digital presence, we can also help you define a content strategy through our SEO copywriting services, and manage your Google Ads campaigns and LinkedIn campaigns. And if you're considering an upgrade to your web infrastructure, our web services include advice on choosing the technology platforms that fit you best.

To talk through your company's specific needs, you can contact us through our website. And to stay up to date with insights on the digital industry, we recommend following our blog.