State of Enterprise AI SaaS
Where pricing power and AI-disruption resilience are migrating across the enterprise-software stack — US-first, June 2026
Own the consumption-metered data/compute and observability layer; underweight pure seat-based application incumbents until the agentic pricing transition proves net-additive.
As of June 2026 the "SaaSpocalypse" panic has reset (indices back green, IGV still -18% TTM) but the recovery is brutally uneven: the consumption-metered picks-and-shovels — Snowflake (SNOW), Datadog (DDOG), MongoDB (MDB), and private Databricks — are getting paid because agent deployment grows their bill, while seat-based incumbents Salesforce (CRM) and ServiceNow (NOW) must prove that agentic SKUs (Agentforce, Now Assist) add net revenue faster than agents compress seat counts. Own the data/observability layer and a diversified IGV core; treat CRM/NOW as show-me re-rate candidates (cheap optionality at CRM, premium-priced execution at NOW) and PLTR as a high-conviction-narrative / extreme-multiple (~200x) name to size small; for broad exposure use IGV (software) and WCLD (cloud) rather than single-name seat-based bets. Not financial advice.
No single layer dominates the way a chip-stack bottleneck does — but the consumption/usage-metered data + compute + observability layer is the clearest structural winner of the AI movement, because its revenue is mechanically LONG agent adoption while the seat-based app layer is mechanically SHORT it. The deepest moats are the systems of record (data gravity, switching cost, compliance) — but the value is leaking from the workflow ABOVE them, not the record itself. Pure pricing power as of June 2026 sits with consumption-metered infra that meters the agent buildout, not with the application brands that historically captured the enterprise budget.
Full-stack value chain
Each layer scored on whether the AI movement is strengthening or eroding its pricing power and disruption-resilience — June 2026
The AI movement is re-pricing enterprise software, not ending it. Pricing power and disruption-resilience are draining out of the seat-priced application layer (CRM, NOW — mechanically SHORT agent adoption) and pooling in the consumption-metered data, compute, and observability layers (SNOW, DDOG, MDB, Databricks — mechanically LONG it), because every deployed agent queries more data, emits more telemetry, and burns more compute but buys zero new seats. The catch: the winning layer trades 80-90% seat-economics gross margins for 50-65% AI-native margins (inference is a real 'token tax'), so the structural long on growth is structurally lower on margin quality. Own the consumption/observability layer and the systems of record with proprietary data moats; treat the seat-based incumbents as show-me re-rate candidates whose agentic SKUs must prove net-additive; use IGV/WCLD for diversified core rather than concentrating in single seat names. Not financial advice.
Upgrade to Pro to see the full value chain.
Opportunity board
Where to lean in, what to avoid, and the ETF routes — ordered by conviction in the durable economics, not the hype
Durable compounder
5Undervalued / high-potential
4Short / avoid
0No names in this bucket.
Upgrade to Pro to see the full opportunity board.
Full board analytics
The prediction matrix, scenario bands, regime timeline, book construction, premise tests and the shift-point register — the deep analytical layer behind this board.
- Prediction matrix & scenario bands
- Book construction & sizing
- Premise tests & falsifiers
- Shift-point register