Jason Lemkin's 'Tired vs. Wired' Framework: Four Categories, and 'The Recovery Is Not Coming for Category Four'

SaaStr·high signal

Lemkin resolves the paradox of software spend growing 15% to $1.4T while software stocks trade at S&P discounts by splitting vendors into four buckets: AI-native with no legacy constraints (Harvey, Artisan, Monaco); pre-AI companies catching AI tailwinds (Twilio, WorkOS, RevenueCat); AI-driven expansion only, no new logos (Atlassian); and no AI advantage at all. Palantir accelerated 27% → 85% growth, Twilio rebounded ~4% → 20%. SaaStr itself went from 20+ employees to 3 humans plus 21 AI agents. The useful diagnostic for any vendor: is AI bringing you new customers, or only bigger invoices from existing ones?