Dmitry Shevelenko: Consumer AI usage has plateaued, revenue is a more reliable metric than user numbers, and the gap between AI capabilities and consumer behavior hinders growth | Big Technology

Dmitry Shevelenko argues that consumer AI adoption has stalled and that revenue metrics are more meaningful than user growth numbers for evaluating AI company performance. He highlights Perplexity's ARR growth as evidence of a widening gap between AI technological capabilities and actual consumer behavior, suggesting this disconnect is the primary constraint limiting mainstream AI adoption.

Shevelenko's commentary addresses a critical tension in the AI industry: sophisticated models and capabilities have advanced rapidly, yet consumer adoption patterns remain sluggish. This observation challenges the common startup metric of vanity metrics like daily active users or sign-ups, proposing instead that annual recurring revenue better reflects genuine market demand and product-market fit. Perplexity's strong ARR performance despite plateau in broader consumer AI usage becomes illustrative—the company demonstrates that focused, revenue-generating products can thrive even when general AI enthusiasm wanes.

This perspective emerges amid broader market skepticism about AI's consumer viability. Many AI startups launched on hype cycles rather than clear value propositions, leading to adoption plateaus once initial novelty wore off. The maturation suggests the market is consolidating around applications with defensible use cases and clear ROI rather than experimental tools.

For investors and developers, this signals a strategic shift: focus on building for specific workflows with measurable outcomes rather than chasing raw user numbers. Companies that monetize effectively through subscriptions or enterprise features outperform those relying on advertising-dependent growth models. The implication extends to capital allocation—VCs increasingly scrutinize unit economics and revenue velocity over user acquisition metrics.

Looking forward, successful AI companies will likely differentiate through specialized capabilities addressing particular industries or professions, rather than horizontal consumer platforms. The gap between capability and behavior suggests consumer education and demonstrated value matter more than raw feature count.

• →Consumer AI usage has plateaued, indicating the novelty cycle has ended and only differentiated products retain user engagement
• →Revenue metrics provide more reliable indicators of success than user counts or vanity metrics in evaluating AI companies
• →The mismatch between advanced AI capabilities and actual consumer adoption patterns constrains market growth
• →Perplexity's ARR growth demonstrates that focused, monetized AI products can succeed despite broader adoption slowdown
• →Capital and development resources should prioritize vertical-specific applications with clear ROI over horizontal consumer platforms