Adoption Curve Compression

In the AI Story

The compression pattern has a specific history. Across the technology waves Meeker has tracked since the 1990s, each new S-curve inherited the distribution infrastructure of its predecessors. The internet reached 50 million users faster than television because television had built the broadcast-reception culture that prepared audiences for rapid media adoption. Mobile reached 50 million faster than the desktop internet because the internet had built the digital content and services that made phones worth carrying.

AI's compression represents a different kind of event. The internet's four years to 50 million users was itself unprecedented at the time. ChatGPT's two months is not a refinement of that pattern but a break from it. The technology achieved mass adoption not through progressive infrastructure expansion but through immediate access to infrastructure that already existed at saturation scale.

The mechanism has quantitative consequences that Meeker's 2025 report documents in detail. Because AI inherited existing infrastructure rather than building new, the capital structure of the transition differs from prior waves. The Big Six technology companies invested $212 billion in AI infrastructure in 2024 alone — capital flowing not into speculative ventures but into companies with existing revenue streams, existing distribution networks, existing customer relationships.

The compression has cultural implications that exceed the data. When a technology reaches mass adoption in two months, the institutional response — educational, organizational, regulatory — cannot keep pace. The institutional lag that characterized prior technology transitions lasted years; the lag for AI is measured against a timeline that institutions, by their nature, cannot match.

Origin

The pattern was first identified across Meeker's Internet Trends reports of the 1990s and 2000s, which routinely stacked adoption curves to reveal the progressive compression across successive technologies. The analytical technique — overlaying curves to reveal structural parallels — became one of the signatures of Meeker's method.

The 2025 report applied the stacking technique to AI and produced a chart that has since been reproduced across the technology discourse: every prior S-curve, plotted against ChatGPT's trajectory, reduced to near-horizontal lines by comparison.

Key Ideas

Compression is the pattern. Each general-purpose technology reaches mass adoption faster than its predecessor, and the acceleration itself is accelerating.

Step functions, not curves. AI's adoption does not refine the pattern; it breaks it. Two months to 100 million is not a faster version of four years; it is a different kind of event.

Compounding inheritance. Each new technology inherits the infrastructure of prior technologies. AI inherited all of them simultaneously — internet, mobile, cloud — producing unprecedented speed.

The institutional gap widens. When technology compresses to months and institutions adapt over years, the gap between capability and governance becomes structural.

Compression is not capability. The curve measures how fast the technology spreads. It does not measure whether the spreading produces value commensurate with the speed.