Korva is a scroll of the internet’s best ideas.
$9 billion to $30 billion in months
The jump in numbers
The company is only a little over five years old. This past March was the third anniversary of its first dollar of revenue, and it had a frontier model for real for the first time in March of 2024. The growth came from leaps in model intelligence and the products built around them.
What the doubters thought
At the end of 2024, Anthropic raised its Series E with run rate revenue close to a billion dollars. Some people still had questions. They looked at the forecast, saw another 10x coming, and doubted it was possible: an enterprise business was great, but they figured adoption would take much longer, and there were laws of physics on how fast a company could scale.
Model led growth
What happened instead was model led growth, carried by the products, the go-to-market team, and distribution. When one person joined, run rate revenue was about $250 million and the plan was to reach a billion. The first question was in what year, which is linear thinking, and the thinking has since shifted toward leaning into the exponential.
Why does run rate revenue move so fast?
Run rate takes the revenue of a recent period, often a month or a single week, and multiplies it up to a full year. That means a company with usage-based pricing can see its run rate jump the moment a few large customers ramp up their spending. It is a forecast of the year if today's pace held, not money already collected. Startups and investors favor it because it reflects current momentum, while trailing annual revenue lags months behind.
Why do enterprise sales usually take so long?
Large companies typically move a new software tool through security review, legal negotiation, procurement, and a pilot before wide rollout. Each step involves a different team with its own sign-off, so a deal can take many months from first call to full deployment. Products that individual employees or developers can start using on their own, with a credit card or an API key, skip much of that queue. That is why adoption of tools sold through developer access can outrun the timelines that traditional enterprise software follows.
Why do forecasts struggle with exponential growth?
People tend to extend a trend as a straight line, adding the same amount each period, while exponential growth multiplies by the same factor each period. The old legend of the chessboard and the grains of rice shows the gap: doubling a grain across 64 squares ends in a quantity larger than any harvest. Moore's law, the observation by Intel co-founder Gordon Moore that chip transistor counts doubled roughly every two years, kept surprising planners for decades for the same reason. Early points on an exponential curve look small, so a straight-line forecast fits them well until the curve bends upward.
Why does growth usually slow as companies get big?
A company's growth rate is measured against its existing base, so the same dollar gain counts for less each year. Going from $1 million to $10 million takes ten times the revenue, and going from $10 billion to $100 billion takes ten times again, but in far larger absolute amounts. Customers, sales staff, and available buyers in a market also run into limits that small companies never touch. Analysts often call this the law of large numbers in business, and it is why sustained 10x years are rare at scale.