How The Sector Breaks Down

Seventeen companies, sorted into three competitive positions plus a no-show top tier. No company has yet crossed the bar into Dominant Innovator. Seven are Disruptive Challengers — the AI-native IDEs, autonomous coding agents and prompt-to-app builders carrying category-defining momentum. Nine are Emerging Players executing in lane. One is Wound-Down. The coding-AI Compass picture is structurally distinct from legal AI (two Dominant Innovators) or foundation-model providers (six): the cohort is being squeezed from above by the foundation-model providers themselves, and the moat geography is still resolving.

The Information Matters Compass — Coding AI Sector

The Information Matters Compass plots every covered company on two axes — how defensible the business looks (left–right) and how much disruption potential it carries (bottom–top). The dashed lines at 7.5 split the chart into four equal quadrants. The coding-AI cohort sits almost entirely below the Defensibility bar: switching costs are low (the IDE-and-model substrate is interchangeable), proprietary data moats are thin (public-code corpora are commoditised), and the foundation-model providers are entering the same surface from above. Disruption Potential is where the variation runs — led by Anysphere, Cognition Labs, Replit, Bolt and Lovable.

The 17 Companies

Defensibility and Disruption are scored 0–10; Overall is the weighted combination. The numbers are Information Matters’ assessments, applied consistently across the cohort, and audited before publication.

What This Tells Us About Coding AI In 2026

The seventeen coding-AI companies split into three distinct shapes, with the top-right quadrant conspicuously empty. No company has reached the top-right Dominant Innovator quadrant. GitHub Copilot is the closest entrant on Defensibility (composite just over 7.1, on the strength of Microsoft-owned distribution and VS Code integration) but sits just over the 7.5 bar on Disruption Potential rather than well above it. Anysphere carries the highest Disruption score in the cohort — a 9.1 composite on the back of a $0-to-$500M+ ARR ramp in roughly twenty-four months, the fastest revenue trajectory in the entire IM tracker — but Defensibility sits in the low sixes because switching costs against Claude Code, Codex and Gemini Code Assist are structurally thin. The pattern is the central finding of Information Matters Category Report #IM107 and Macro Briefing #IM104: coding AI has produced the fastest revenue ramps in the agentic-AI cycle, but it has not yet produced an entrant that compounds those ramps into top-tier defensibility.

Seven Disruptive Challengers cluster top-left. Anysphere, GitHub Copilot, v0/Vercel, Cognition Labs, Replit, Bolt/StackBlitz and Lovable all carry Disruption scores above the 7.5 bar, but each sits below the Defensibility line. The sub-segment split is sharp. Anysphere and GitHub Copilot lead the AI-native IDE sub-segment, on the strength of Cursor’s hyper-growth ramp and Copilot’s Microsoft-distribution moat respectively. Cognition Labs leads the autonomous-coding-agent sub-segment with Devin, on the basis of end-to-end-engineer positioning and the highest Disruption score after Anysphere. Replit, Bolt and Lovable lead the prompt-to-app sub-segment — Replit on the strength of incumbent browser-IDE distribution, Bolt and Lovable on rapid prosumer traction. v0 sits in a sub-segment of one, the generative-UI specialism inside Vercel’s deployment platform.

Nine Emerging Players sit mid-table. CodeRabbit, JetBrains AI, Augment Code, Continue, Tabnine, Tabby, Sourcegraph, Magic and Aider all score in the four-to-seven range on both axes. CodeRabbit is the closest to the Disruption bar on the strength of category-defining PR-review specialism. JetBrains AI carries the strongest Defensibility of the group on the back of incumbent IDE distribution. Augment Code carries enterprise-grade context-engine differentiation. Sourcegraph (Cody) layers AI on top of the code-search incumbent. Continue, Tabby and Aider hold the open-source positioning. Tabnine carries the longest enterprise and on-prem deployment book. Magic sits at the smaller-cap end on long-context-model research positioning.

One Wound-Down residual. Sweep is the realised exit from the cohort — an early autonomous-coding-agent project that did not compound and entered operational wind-down. The Wound-Down flag is operational status, not a Compass position; Sweep is shown for completeness with its true scores noted.

The margin-reset thesis is the structural overlay. Information Matters Macro Briefing #IM104 sets out the central finding: the coding-AI cohort is being commoditised from above by the foundation-model providers. Anthropic’s Claude Code, OpenAI’s Codex and Operator, and Google’s Gemini Code Assist now ship coding-agent capabilities directly inside the model-provider product surface, at zero marginal distribution cost. The economic consequence is structural margin compression on the AI-native IDE and agent layer: the value the cohort captured in 2024–2025 (model orchestration, IDE-native context, agentic loops) is being absorbed into the model-provider product itself. The Compass picture above is the snapshot of that compression in progress — high Disruption scores on the back of category-defining momentum, low Defensibility scores on the back of the same substrate becoming free.

The Workers future is the structural overlay. The Information Matters Framework names eight futures — the patterns the next ten years of AI will resolve into. The coding-AI cohort is the clearest expression of the Workers future of the eight. Every vendor in this cohort sits inside a buyer-side process where the unit being automated is professional software-engineering labour. Power shows up too, in the foundation-model-provider margin pressure that drives the #IM104 thesis — the providers with the largest training-compute book are the same providers entering the agent layer from above. The Compass picture makes both legible.

The deep dive. This sector page is the higher-level Compass entry point on the IM-Framework-scored subset. The full qualitative competitive treatment, the sub-segment maps and the vendor profiles live in Information Matters Category Report #IM107 — AI Coding Agents: The 2026 Landscape. The structural margin-reset thesis — how foundation-model providers reshape the cohort over the next 12–18 months — lives in Information Matters Macro Briefing #IM104 — Coding agents: the margin reset is coming.

How To Read These Scores

Every company is scored on nine plain-English dimensions. Defensibility covers how sticky the customers are, what proprietary knowledge or data the company holds, the strength of its distribution channels, its strategic resilience to shocks, and whether it benefits from platform-style network effects. Disruption Potential covers momentum, how novel the capability is, how fast the team executes, and how much category leadership the company commands. Each dimension is scored from 0 to 10. A sector-appropriate weighting produces the Defensibility and Disruption composites that drive the Compass position.

The competitive position labels — Dominant Innovator, Disruptive Challenger, Established Incumbent, Emerging Player — come from where the composites place a company on the Compass, not from analyst judgment. A separate Wound-Down label is used for residual entities post-acquisition or wind-down; one vendor in this cohort (Sweep) carries that status. For the full methodology, including how each dimension is broken down further, see the Information Matters Framework Scoring methodology. Every score on this page has been through Information Matters’ two-layer audit before publication.