It's Accelerating Growth: the Second Derivative Is Positive; but "ARR" Hides Three Different Things — A July 8 TechCrunch piece lines up the revenue curves and the common thread isn't "fast," it's accelerating: Sierra took 7 quarters to reach its first $100M ARR and only 2 more for the next $100M; Glean went $100M→$200M in 9 months, then $200M→$300M in 6; Anthropic went $30B→$47B run-rate in under two months; Mercor $1B→$2B in 4 months; Clio and Gusto also accelerating quarter over quarter. Sources: TechCrunch / company disclosures. Judgment: (1) classic SaaS law says growth decays with scale — a bigger base dilutes the same absolute gain into a lower percentage; these companies invert it, scaling while the second derivative stays positive, because consumption billing + agents lift software's unit value from "save someone effort" to "do the work," moving budget from IT to headcount/business lines and raising the ceiling by orders of magnitude; (2) the trap: when everyone reports "ARR" but underneath sits recurring vs run-rate (annualize the best month) vs committed (signed, not onboarded), the number slides from accounting to marketing — run-rate has the most fragile denominator because AI consumption can churn as fast as it arrives; ask the definition, ask net retention, ask if the revenue is still there next month; (3) primary-market read: absolute value stops being the signal — the second derivative is — but separate real acceleration (retention and margin move with it) from definitional acceleration (committed stuffed into run-rate); the growth anchor has moved, $100M ARR is no longer remarkable, the market now asks "how long for your second $100M." Rollback (ties to ENTRY 95's meter): a positive second derivative cuts both ways — it turns negative just as sharply, and run-rate pricing falls faster than recurring once enterprises cap consumption; the prettiest curve is the one most in need of falsifying against retention and margin — growth lies, retention doesn't.