Annualized Run Rate

Annualized Run Rate

The Annualized Run Rate projects the revenue of a short period onto a full year. It shows how much a company would earn if the current pace continued unchanged for twelve months.

The Annualized Run Rate is a projection of revenue. You take what a company earned in a short period and extrapolate it to twelve months. Anyone earning 10 million euros in a month has an Annualized Run Rate of 120 million euros. What matters is this: this money hasn’t actually come in yet. The figure only describes what would result if the current pace stayed unchanged for a year. That’s why it’s a snapshot, not an annual result.

Why young companies love calculating it

A normal annual revenue figure looks backward. It sums up twelve months that have already passed. For a company changing rapidly, that’s a poor description. A company that earned almost nothing in January and 20 million in December looks small in a year-end review. Its Run Rate, by contrast, shows today’s standing.

That’s exactly why this metric is ubiquitous in the AI industry. Companies like OpenAI or Anthropic grow in months, not years. Investors and journalists want to know where a company stands right now. The Annualized Run Rate delivers a single figure that can easily be compared to established corporations.

But that same property also makes it seductive. Anyone who chooses the best month of the year as a basis produces a particularly large number. Nobody forbids this, because the Annualized Run Rate is not an audited accounting metric. It shows up in press releases and investor conversations, not in official annual reports.

The calculation behind it

The formula is simple. You take the revenue of a period and multiply it so that a year results. A month is multiplied by twelve, a quarter by four. 300 million euros in quarterly revenue thus becomes 1.2 billion euros in Annualized Run Rate.

The catch lies in the underlying assumption. The calculation presumes that nothing changes: no new customers, no cancellations, no fluctuations. In reality, this is almost never the case. A ski rental company with a February run rate of 12 million euros will see very different numbers in July.

It’s also important to note: the figure says nothing about profit. It describes revenue, not what’s left over in the end. An AI provider can report a run rate of one billion and simultaneously post losses because data centers and chips cost even more. Both statements can be true at the same time without contradicting each other.

How to read the figure in headlines

In tech news, you almost always encounter this metric in the same form. A report might say an AI company has reached an Annualized Run Rate of several billion dollars. Often only ARR appears there. This abbreviation is also used for subscription companies to mean Annual Recurring Revenue, i.e., the recurring yearly income from ongoing contracts. That’s something different and considerably more reliable.

Two questions are worth asking when reading such figures. First: which period was extrapolated? A single strong month is weaker evidence than a full quarter. Second: is this recurring revenue or one-off deals? A single large one-time order can massively distort the projection.

Still, the figure is useful, especially over time. If a company’s run rate triples within a year, that’s a genuine signal. A common mistake is simply confusing it with actual annual revenue. The run rate is a forecast based on the present, not an account balance.

Related Products

Latest News

Subscribe free. Unsubscribe the second it sucks.

High-signal news across AI, business, UX, and tech. Every morning.