Revenue Run Rate

A company takes in money from its customers over the course of a month. This income from selling products or services is called revenue. The revenue run rate extrapolates this revenue to a full year. If a company earns 10 million euros in December, its run rate comes out to 120 million euros per year. What matters is this: the company never actually took in those 120 million. It’s an estimate that assumes every following month performs just as well as December.

Why young tech companies love calculating in run rate

Traditional annual financial statements look backward. They report what happened over the past twelve months. For a fast-growing company, that gives a distorted picture. If an AI company brings in 2 million euros in revenue in January and 10 million in December, its annual revenue might come out to 60 million. The run rate at year’s end, on the other hand, says 120 million. Both figures are correct, but they answer different questions.

For investors, the current pace is often more interesting than the past. They’re buying a stake in a company’s future, not its history. That’s why run rate figures constantly show up in funding rounds and press releases. OpenAI, Anthropic, and other AI providers report their progress almost always in this form.

But that is precisely where the danger lies. The run rate is not an audited figure from a financial report. Companies are largely free to calculate and publish it however they like. Anyone who takes the best month of the year as their basis ends up with a flattering number. A critical reader therefore always asks which time period was used for the extrapolation.

Extrapolating from one month to twelve

The calculation itself is simple. You take one month’s revenue and multiply it by twelve. Alternatively, you take a quarter, i.e. three months, and multiply by four. Some companies even take just the last week and multiply by 52. The shorter the period, the more strongly random fluctuations show through.

The method only works under one silent assumption: that business continues at a steady pace. For a software subscription, that’s often realistic, since customers pay monthly. For a toy retailer, it would be nonsensical. Its December revenue times twelve would produce a fantasy figure, because the holiday shopping season is a one-time event. Such seasonal effects are the most common reason run rates lead people astray.

A related term is ARR, Annual Recurring Revenue. It counts exclusively recurring income from ongoing subscriptions. One-time sales or consulting fees are excluded. The revenue run rate is broader in scope and includes everything that has come in recently. ARR is therefore considered the tougher, more reliable metric.

The number behind the headlines about AI companies

When you read in business news that an AI startup has “reached a billion dollars in revenue,” there’s almost always a run rate behind it. The company actually took in around 83 million dollars in a single month. The distinction sounds like nitpicking, but it’s substantial. When in doubt, it’s worth checking the original report.

The metric is also common within companies. Management teams use it to plan budgets or set targets. A sales team, for instance, might be given the goal of doubling the run rate by year’s end. That’s more tangible than an abstract annual forecast.

For you as a reader, one question is especially useful: is the run rate growing across multiple reports? A single figure says little. A series of figures, however, shows whether growth is continuing or leveling off. It’s precisely this trajectory that often determines a company’s value on the stock market.