S-curve

The S-curve is a diagram that shows a development over time. Its name comes from its shape: the line looks like an elongated S. At the beginning it barely rises, then it shoots up steeply, and toward the end it flattens out again. A good example is smartphones: after 2007, only a few people owned one at first, then almost everyone bought one, and today in Germany hardly anyone is without one. The market is saturated, further growth is barely possible. Exactly this pattern repeats itself with very many technologies.

Why forecasts so often miss the mark

People like to think in straight lines. If something has grown at the same pace for three years, we expect the same pace for the next three years. The S-curve shows why this goes wrong. In the flat initial phase, a technology is almost always underestimated because it seems so insignificant. In the steep phase, it is overestimated because the later flattening is not factored in.

For investors and companies, this has real consequences. Anyone who buys a product during the steep phase often pays prices that assume eternal growth. But such growth does not exist, because at some point all possible customers have been reached. Conversely, companies miss their entry point if they write off a technology during the dull initial phase. Kodak, for example, long considered the digital camera a mere gimmick.

Around artificial intelligence, the S-curve is therefore constantly used as an argument. Optimists say: we are only at the beginning of the steep part. Skeptics say: the steep part is almost over, the flattening is coming soon. Both use the same curve and arrive at opposite conclusions. Which side is right is always known for certain only in hindsight.

The three phases and their brakes

The first phase is the ramp-up phase. A new technology is expensive, unreliable, and of interest only to a few specialists. There is a lack of accessories, knowledge, and trust. This is why almost nothing grows here, even though the invention basically already works.

Then it tips over. The technology becomes cheaper, imitators enter the market, and every new user indirectly recruits further users. This self-reinforcing effect produces the steep rise. In mathematics, such a course is called logistic growth. It also describes how diseases spread within a population.

In the end, something always slows it down. For products, it’s the market: at some point all possible buyers have been supplied. For technologies, it’s physical limits, raw materials, or costs that rise faster than the benefit. It’s important to distinguish this from exponential growth. Exponential means: it keeps getting steeper forever. An S-curve looks the same at the beginning, but it has a built-in upper limit.

S-curves in stock market reports and tech news

In quarterly reports of technology companies, this pattern often appears without being named. When a company reports that user growth is slowing, it is describing the upper end of an S-curve. The stock price often reacts sharply to this, even though the company is still growing. What was expected was the steep part, what was delivered was the flat part.

There is also debate about this with AI models. Every new generation is supposed to be significantly better than the previous one. However, some researchers observe that the gains are getting smaller, even though computing power and data volumes are increasing sharply. This would be a sign of flattening. Others counter that a new idea can start a second S-curve at any time.

This stringing-together is the single most important idea. Progress as a whole consists of many S-curves lined up one after another. Carriages were replaced by cars, tube televisions by flat screens, CDs by streaming. Each individual curve ends, but the next one often already begins before the old one runs out. Anyone who looks only at a single curve therefore quickly sees an ending, where in reality a transition is taking place.