How VCs and founders use inflated ‘ARR’ to crown AI startups

There's a quiet trend happening in AI startup land: companies are redefining what counts as ARR (annual recurring revenue) in ways that would make traditional SaaS founders raise an eyebrow.

The classic definition is pretty straightforward. Take your monthly recurring revenue, multiply by 12, and that's your ARR. But some AI companies are now including one-time deals, pilot programs, or usage-based revenue that might not actually recur.

Here's the thing: their investors aren't being fooled. VCs backing these companies are fully aware of the creative accounting. In some cases, they're even encouraging it as a way to build momentum and attract the next round of funding.

Why does this matter if you're building with AI? Because these inflated numbers set unrealistic benchmarks for what success looks like. When you see a competitor claiming massive ARR growth, there's a decent chance they're measuring it differently than you think.

It also affects how the market values AI companies overall. If enough startups use loose definitions, it becomes harder to tell which ones have genuinely sticky revenue versus which ones are riding a wave of experimental budgets.

The practice isn't necessarily fraudulent, but it does make due diligence more important. Whether you're evaluating tools to buy, companies to join, or startups to invest in, ask how they're actually defining their recurring revenue.