Where peers are moving.

A peer team gets a renewal quote from Splunk, opens a parallel evaluation of Cribl in front of Snowflake, and within two quarters the observability spend has been cut in half while the data the SRE team actually wants now lands in the warehouse it already pays for. That single arc, repeated with different names in different categories, is the pattern this issue reads across 163,000+ organizations and 127,000+ deployment case studies in the Airframe registry. Legacy stack on the left, modern stack on the right, twelve named paths in between. The diagrams are illustrative of pattern and direction in the corpus, not measured volume counts.

The read

Three signals, every time.

The Airframe corpus has been picking up the same three signals ahead of every named migration in this report. They show up in the deployment narratives 18 to 36 months before the incumbent posts the quarter that confirms it. They are not predictions, they are the cadence the registry surfaces.

01The incumbent’s product stops moving.
02The challenger wins net-new projects.
03The economics invert.

Fig. 01 · Legacy stack to modern stackDirectional · patterns in the registryRibbon thickness is schematic, not a measured countIllustrative of patterns the registry surfaces

The incumbent is not losing because the insurgent is better. The incumbent is losing because the category rewrote itself around them.

Airframe Research

Twelve paths, named.

The twelve legacy-to-modern paths the registry surfaces most often, written with the incumbent and challenger named in each row. Bars are editorial weight in the corpus, not a measured ranking; reason codes are the dominant driver read in each case, aggregated from deployment narratives across the Airframe corpus.

01Splunk→Cribl + SnowflakeBill shock, route the noise, land the rest in the warehouse already paid for

02Confluence→NotionAI-era surface, blocks model, search and assistants the writers actually use

03Jenkins→GitHub ActionsArchitect preference, config-as-code, no plugin sprawl, runners where the code lives

04Legacy code editors→CursorAI-era capability gap, editor-native agent, multi-file edits, the IDE became the surface

05Teradata→SnowflakeCost compression, separation of storage and compute, no more capacity ceiling

06Cloudera→DatabricksAI workload gravity, lakehouse as the ML-native surface, unified governance pitch

07Tableau / Looker→HexNotebook-native, SQL plus Python in one cell, analyst workflow rewritten around AI

08Siebel→SalesforceWave II displacement, multi-tenant cloud finally retiring the last on-prem CRM seats

09PeopleSoft→WorkdayConsolidation play, HCM and finance on one ledger, mobile-first manager workflows

10Ariba→Coupa + RampContract and leverage cycle, procure-to-pay finally splits from card and expense controls

11Legacy enterprise search→GleanAI-era capability gap, retrieval over every SaaS surface, answers instead of links

12Manual GTM tooling→ClayAI-era capability gap, enrichment plus orchestration, agents wired straight into outbound

Fig. 02 · Twelve named legacy-to-modern pathsBars are editorial weight, not a measured countReason codes aggregated from deployment narrativesDirectional read of patterns the registry surfaces

Why they moved.

The ten reason codes the registry most frequently surfaces as the primary driver, read across 127,000+ deployment case studies. Bars are editorial weight, not precise percentages.

01Bill shock, cost compressionDominant

02AI-era capability gapStrong

03OSS and self-host leverageStrong

04Architect and SRE preferenceNotable

05Consolidation playNotable

06Contract and leverage cycleNotable

07Governance and data residencyEmerging

08Acquired vendor, forced moveEmerging

09Talent and hiring gravityLong tail

10Other, idiosyncraticLong tail

Fig. 03 · Reason codes the registry surfaces most oftenEditorial weight, not precise percentagesBars directional, relative to the dominant patternIllustrative of patterns the registry surfaces

Bill shock is the loudest single driver, but the joint weight of the next three codes (AI-era capability gap, OSS leverage, architect preference) reads as substantially larger than price alone, and that is the signal worth holding onto. When a category moves, the defense almost always has to move with it rather than against it, which is why the most durable reads in the corpus describe the new surface the work has shifted onto, not the old line item the bill used to sit on.

The migration the corpus is accelerating into Q3 is the legacy-editor exit (path 04, Cursor and the editor-native agents) running into the legacy-enterprise- search exit (path 11, Glean and the retrieval layer) inside the same engineering organizations, on overlapping renewal windows. When those two arcs hit the same CFO desk in the same quarter, the conversation stops being a renewal and starts being a workflow review, and that is when the next named path in this report tends to open up. The twelve rows above are a starting point for your own peer conversations in the registry, not a substitute for them.

The registry behind the read

An Airframe editorial pattern is written when the corpus repeatedly surfaces directional movement across independent deployment artifacts. To see the underlying registry, talk to us at [email protected].

Organizations

163,000+ tracked in the registry

Signal base

127,000+ deployment case studies

Vendor corpus

2,406 software leaders, 1969 to 2026

Independence

Zero vendor money behind the registry. We are paid the same whichever tool the read points to.

Talk to us

Bring this audit to your stack.

Airframe maps your registry, your tools, your owners, your spend, against the cohort. Thirty minutes to scope.

Airframe · 2026

Coverage19,000+ tools · 163,000+ organizations · 127,000+ deployment case studies.

IndependenceZero vendor money behind the registry. We are paid the same whichever regardless of which tool the read recommends.

Contact[email protected]