Technology Adoption S-Curve
Technology Adoption S-Curve
One-line summary: New technologies are adopted along an S-shaped (logistic) curve — slow at first, a non-linear acceleration past a tipping point, then deceleration toward a saturation ceiling — which is the single most robust stylized fact in the study of technological change.
The insight
Adoption over time is not linear; it traces a sigmoid. RethinkX frames the phases as: an initial phase where growth looks slow because penetration is under 1-2% ("an illusion of gradual change"), a tipping/"rupture" point where "change is almost inevitable," and a saturation phase near ~80%. The qualitative point: a disruption is a phase change, not a faster incumbent — "A butterfly is not a faster caterpillar." Critically, the same S-curve can be generated by several distinct mechanisms — information epidemics (limited by how fast knowledge of the technology spreads), firm/agent heterogeneity (different adopters cross their threshold at different times), and population-ecology density dependence — which matters because each implies a different lever on speed.
The mathematical model is the logistic y = L / (1 + e^(−k(x−x₀))), with three parameters that map onto the three investing questions: L = ceiling/market potential ("how big?"), k = steepness ("how violent the re-rate?", see curve-steepness-and-adoption-velocity), and x₀ = inflection/midpoint ("are we before or after the knee?").
Evidence
- From 2026-06-16-academic-research-technology-adoption-s-curves: the S-curve is "the dominant stylized fact: that the usage of new technologies over time typically follows an S-curve" (Geroski 2000), arising from epidemic, probit (firm-heterogeneity), and density-dependence mechanisms.
- From 2026-06-16-autoresearch-tony-seba-technology-disruption-s-curves: "The adoption of the new technology is non linear and follows an S-curve … slow at first because a new product has less than 1-2% market penetration, and it then hits a tipping point and accelerates until the product nears about 80% of the market" (RethinkX).
- From 2026-06-16-autoresearch-tony-seba-technology-disruption-s-curves: the logistic model y = L/(1+e^(−k(x−x₀))) with parameters L (ceiling), k (steepness), x₀ (inflection) (Dr Li).
Design implications
- The curve has to be read per technology and per segment, not per theme — within one "energy transition," solar can be beating forecasts while EV battery demand undershoots (see adoption-underestimated-or-overhyped).
- Where you enter the curve determines how much growth is left — the investing core, see s-curve-position-in-stock-evaluation.
Contradictions / tensions
- Real curves are frequently asymmetric around the inflection, so a symmetric-logistic fit can misplace "how much is left" (Easingwood et al. 1981, via 2026-06-16-academic-research-technology-adoption-s-curves).
- The ceiling L and the inflection x₀ are the hardest parameters to estimate prospectively — see detecting-s-curve-inflection-in-real-time.
Open questions
The chain
Falling unit cost (see wrights-law) crosses a price/performance tipping point → adoption goes non-linear along this S-curve → linear-extrapolating forecasters underestimate it. Canonical: cost-curve-tipping-point-to-s-curve-adoption.