In category after category, the registry keeps surfacing more vendors and less consolidation, and what once looked like a temporary Cambrian moment has settled into the default state of enterprise software now, which means the question leaders face is no longer which vendor to pick but which fracture they are buying into.
Vendors in the enterprise software registry, 1974 through 2026. Every legacy category we read in that ledger has more vendors now than in the wave before. None consolidated.
The four waves the registry has lived through: mainframe to client-server, on-prem to SaaS, SaaS to cloud-native, and the one underway now, cloud-native to AI-native. The curves keep climbing.
Tools currently tracked in the Airframe registry, across every category where AI is reshaping the buy. The number grew last quarter. It will grow next quarter.
Peer deployment case studies indexed in the Airframe corpus, the backbone of every category trajectory shown in this issue.
Twelve categories, read across the registry's rolling window. The charts are indexed rather than absolute, because the shape of the fracture is the story the curves tell, not any one endpoint. Read left to right: every category moves up and to the right, and most of them break away from the ink line once the AI wave begins to pull on them.
When a category doubles, that is a competitive moment, a handful of insurgents showing up to challenge a couple of incumbents. When a category multiplies many times over in the same window, the word competition stops being useful, because no procurement process can reasonably evaluate that many options at once. That is the moment the category has fractured, and it is a pattern the registry keeps surfacing in category after category.
The grid below shows every tracked legacy category behaving the same way in roughly the same window. Observability fractured first, because infrastructure surface expanded first. Data warehousing followed. Then analytics, workflow, customer support, content, and code, in that order, within a narrow stretch of one another. The categories that look last to fracture are not the ones the registry expects to consolidate. They are simply the ones whose curves have not yet bent.
There is no "market leader" in a fractured category. There is only which fracture you are inside.
A category that did not meaningfully exist a few years ago now contains many funded vendors, each positioned on a slightly different axis of autonomy, integration, and trust. The landscape below plots the category along the two axes that matter most to a CIO reading the registry today.
Not long ago, code copilot meant one product. A few releases later it meant a handful. Today the category contains many venture-backed vendors alongside an increasingly crowded open-source flank, each claiming a defensible position. The reason this category fractured faster than any other is that the underlying capability is broadly accessible. A plausible copilot can be assembled from foundation-model APIs in a single quarter, which means every credible engineering team has at some point considered shipping one.
For a buyer, the practical question is not "which tool leads the category," because the category does not have a leader in any meaningful sense. The practical question is where on the landscape a given tool sits. The two axes that differentiate these tools, developer autonomy and IDE integration depth, predict almost every downstream decision, from licensing to security posture to the team shape required to adopt them.
The hot zone, editor-native and high-autonomy, is the corner where the case studies keep surfacing net-new enterprise spend, and it is also the quadrant where the category is most unsettled, because every vendor in it is still rewriting their own trust model in public. The remote-agents quadrant is younger and wilder, and a pattern the registry continues to watch closely. The ambient-helpers quadrant is the one where most incumbents still live, and where the quiet attrition the registry keeps surfacing is happening.
Read against the deployment corpus, the fight in this category has already moved past inline completion and chat-with-codebase, both of which are now effectively table stakes. The capabilities that still differentiate are autonomous task execution, codebase-wide indexing and memory, long-horizon refactoring, policy-driven tool scoping, and full prompt-and-output audit. Vendors pitching wave-one features at wave-two prices are the ones the registry sees displaced first, usually within a quarter or two of a wave-two entrant landing inside the same engineering org.
None of this is a ranking. It is a map. The argument of this dossier is that maps are what a fractured category rewards, and rankings are what it punishes.
The index below groups the tracked categories by the shape of their climbing curves in the registry. The ordering is directional and reflects the Airframe registry's tracked footprint within each category. It is not a market share claim.
The Airframe registry is an AI-maintained index of tools in categories where AI is reshaping the buy, grounded in 1,433 vendors tracked from 1974 through 2026 and 114,000+ peer deployment case studies. Category assignment is multi-label. Vendor counts in this issue are the tracked footprint: a vendor is included in a category when the Airframe pipeline has ingested at least one deployment signal, funding record, or product artifact attesting active status in that category in the given year. Figures are directional, not audited.
A small group of Fortune 1000 organizations is building their AI system of record with us now. Free to start. No vendor sponsorship. Pre-launch terms carry forward beyond launch.
Apply to the cohort →