Curves, Not Lines

Real growth isn't a straight line or an endless rocket — it's an S-curve, and knowing which phase you're in beats every guess.

The Future of Technology · Lesson 3 · 7 min read

A technology's sales have doubled every year for five years. One analyst draws the line straight up and forecasts that in a decade it will sell more units than there are people on Earth — obviously absurd. A few years later, when the same curve flattens, a pessimist points at it and declares, “Growth is over, this whole field is dead.” Both of them just took one slice of the curve and stretched it forever. How do you forecast something whose growth is secretly shaped like an S? Hold that question — by the end you should be able to place any hot technology on its curve and avoid both mistakes.

Every exponential is the bottom of an S-curve

In Lesson 1 we saw that nothing compounds forever — growth eventually meets a limit: everyone who wants one has one, a market saturates, physics pushes back. So the true shape of growth over a full lifetime is not a line and not an endless rocket. It is an S-curve: a slow start, an explosive middle, and a flattening top as it nears its ceiling.

The thrilling exponential phase everyone gets excited about is just the bottom-left of an S you haven't watched bend yet. It feels like forever-growth only because you can't yet see the ceiling it is racing toward.

Most forecasting errors are one phase, stretched forever

The S-curve has three phases, and there is a classic mistake for each:

• Stretch the flat bottom forever → “it will never work” (the skeptic dismissing the early toy).
• Stretch the steep middle forever → “it will grow to infinity” (the forecast of more phones than people).
• Stretch the flat top forever → “growth is over, nothing more will happen here.”

The fix is one question that a straight line never makes you ask: where on the S are we, and how close is the limit? That single question bounds your forecast from both sides.

A household gadget goes from rare to everywhere. Early on: expensive, clunky, slow sales — the flat bottom. Then price drops and it gets genuinely useful: sales explode — the steep middle. Then they flatten — not because it failed, but because nearly everyone who wants one now has one. Forecast from the steep middle and you “predict” more units than humans alive. Forecast from the flat top and you declare the field dead — missing that the next curve is already starting underneath.

Progress stacks S-curves

Here is the grounded-optimism resolution. When one S-curve flattens, a successor technology often starts a fresh S-curve and carries the job further. Vacuum tubes gave way to transistors, which gave way to integrated circuits: each individual curve saturated, yet the envelope of progress kept climbing because a new curve picked up where the last left off.

That is why long-run progress can look exponential even though no single technology grows forever — it is a relay of overlapping S-curves, not one magic line. The honest forecaster watches both at once: this curve's ceiling, and whether a successor curve is already starting beneath it.

You follow a technology that has grown explosively for years and you have to forecast the next decade. Walk the S-curve:
1. Which phase? If adoption is still far from any obvious ceiling, you are likely mid-curve — fast growth can continue for a while, so do NOT dismiss it.
2. Where is the limit? Estimate the ceiling (total possible users, a physical bound). The closer it is, the sooner the curve bends — so do NOT project the steep part to infinity.
3. Is there a successor? If this curve is near its top, look for the new technology starting underneath; that, not the old curve, is where the next growth lives.
Notice that the same model talked you out of both the doom forecast and the hype forecast — and pointed you at the real next question.