Enterprise SaaS

This past earnings season separated winners from losers cleanly. Three traits identify the winners: sells to large businesses (sticky, expensive to switch — SNOW, CRWD, PANW, PLTR, NOW), AI generates NEW revenue not just adoption (SNOW Cortex re-accelerated revenue 30%→34%, Atlassian's Rovo lifted NRR to 120%, Twilio Voice AI), and usage-based pricing capturing the AI volume boom. Score 3/3 → market rewards aggressively. Score 0/3 → market punishes regardless of how much AI is "adopted internally."

Within software, infrastructure-layer companies have the cleanest AI-revenue mechanics. SNOW (data), PLTR (AI ops/orchestration), CRWD, PANW (security) sit close to the base of the stack — more AI agents means more data, more queries, more security alerts, directly more revenue. Contrast with companies trying to use AI for internal efficiency (e.g. WDAY-style margin lift, SHOP-style "AI everywhere internally") — the market discounts cost savings if revenue doesn't grow. The closer you are to the AI base layer, the more mechanically your revenue scales with adoption.

AI capability itself is becoming commodity — anyone can wrap OpenAI or Anthropic, and a year-old feature can be obsoleted by the next model. What survives frontier-model iteration: customer relationships + workflow ownership (SAP's 30-yr ERP grip), proprietary data (CRWD's threat telemetry, NOW's CMDB, ORCL's transactional data), security & compliance (PANW certifications), distribution (MSFT enterprise channel). Be cautious of names whose entire AI story is a single flashy feature.

After IGV's ~40% drawdown, the AI-disruption risk is largely in the multiple, not still ahead. Winners this past earnings season faced the same long-term AI threats as losers — nothing about that risk disappeared — yet they re-rated 25-40% in days when business held up. That's the signature of de-risked expectations: when bars are low, modest positive surprises drive sharp re-rates. Don't look for software companies using AI — look for software companies making money from AI.