Which SaaS survives the agentic shift?
This scorecard ranks public SaaS companies on structural resistance to agentic AI. Every company resolves to one of three routes: full internal replacement, headless survival, or owning the agentic layer. Each factor is scored 1–5; half-points are real. Pillar weights: A ~40% · B ~35% · C ~25%. D1–D2 are a cybersecurity appendix, filled only for PANW and CRWD.
Related research
Jensen Huang politely predicted software companies will become "VARs of AI tokens" — a 90% valuation compression delivered with a smile. As AI collapses the cost of building and maintaining software, the industry's brain moves to the model labs, and everyone else fights over the limbs.
Agentic AI is forcing SaaS down one of three routes: full replacement, headless survival, or owning the agentic layer. Part 1 maps the paths and scores 22 public SaaS names on our VAR-Resistance Scorecard — only two emerge as Agentic Compounders.
Part 1 mapped agentic disruption across SaaS with provisional scores. Part 2 deep-dives PLTR, DDOG, and NET: who owns the agent layer, who is durable plumbing agents pay for, and who is more exposed — so investors can see structural winners more clearly.
Palo Alto is not just another SaaS name under agentic pressure. Sensors, inline enforcement, and identity keep SecOps hard to absorb — while Mythos, platform breadth, and token economics rewrite the bull and the risks. Part 3 goes deep on PANW alone.
How to read this
- Gold row marker = deeper dive — Palantir, Datadog, Cloudflare, Palo Alto Networks.
- D1 / D2 apply only to cybersecurity names (PANW, CRWD). Other names show an em dash, not zero.
- Half-points are real scores.
- Buckets: Agentic Compounders · Outcome-Levered Plumbing · Contested Optionality · Toll Roads · Melting Ice Cubes.
- Click any row for the full factor breakdown. Click column headers to sort. Factor criteria — names, core questions, and the 1–5 scoring rungs — are listed below the heatmap.
Buckets at a glance
The 23-name heatmap
Scores 1 (weak) → 5 (strong). Cell shading interpolates across half-points. Hover a column header for the factor name; full scoring rungs are in Factor criteria below.
Factor criteria
Every factor is scored 1–5 (half-points allowed). Higher is stronger resistance or better agentic positioning. Pillar weights: A structural ~40% · B adaptive ~35% · C valuation ~25%. D1–D2 are a cybersecurity appendix.