ATC SOFTWARE MARKET INTELLIGENCE
ATC combines developer activity, enterprise-adoption signals, product discussion, adoption-observed company breadth, and change over time, then gives Claude or ChatGPT the structured evidence needed to produce a market analysis.
Inspect a real recorded analysis of Cursor and its peers below. New questions require the authenticated MCP connector.
Recorded analysis - no signup to inspect - static data - new queries require MCP
RECORDED MARKET ANALYSIS
This is a recorded run generated from the prompt below using ATC MCP. The public web presentation is adapted for responsive inspection; raw responses and internal evidence identifiers are withheld.
Is Cursor overhyped? Compare Cursor, Claude Code, GitHub Copilot, and Windsurf across developer activity, enterprise-adoption signals, product discussion, adoption-observed company breadth, and time-series momentum. Return a verdict, supporting and conflicting evidence, chart-ready trajectories, developer discussion themes, company evidence, and methodology limits.Monthly developer-activity signal across the AI coding tools in the recorded analysis.
Cursor rises steadily over two years, but Claude Code accelerates after February 2025 and finishes about 2.7 times higher than Cursor by May 2026. GitHub Copilot and Windsurf are flatter by comparison.
Cursor climbed steadily, then decelerated through 2025. Claude Code launched into the measured set in February 2025 and crossed Cursor's raw developer-activity signal around September 2025.
Recent months can lag while public signals are collected and resolved. Missing values are shown as gaps, not interpolated.
Log scale omits zero and missing points from the plotted line.
Source and methodology| Month | Cursor | Claude Code | GitHub Copilot | Windsurf |
|---|---|---|---|---|
| Jun 24 | 1,943 | missing | 780 | 520 |
| Jul 24 | 2,000 | missing | 810 | 540 |
| Aug 24 | 2,200 | missing | 850 | 570 |
| Sep 24 | 2,500 | missing | 890 | 610 |
| Oct 24 | 2,650 | missing | 910 | 650 |
| Nov 24 | 2,800 | missing | 940 | 700 |
| Dec 24 | 2,950 | missing | 960 | 760 |
| Jan 25 | 3,100 | missing | 980 | 820 |
| Feb 25 | 3,300 | 350 | 995 | 850 |
| Mar 25 | 3,450 | 650 | 1,010 | 875 |
| Apr 25 | 3,500 | 920 | 1,025 | 900 |
| May 25 | 3,600 | 1,350 | 1,040 | 930 |
| Jun 25 | 3,700 | 2,050 | 1,060 | 960 |
| Jul 25 | 3,780 | 2,800 | 1,075 | 980 |
| Aug 25 | 3,820 | 3,400 | 1,090 | 1,020 |
| Sep 25 | 3,880 | 3,900 | 1,120 | 1,050 |
| Oct 25 | 3,920 | 5,000 | 1,140 | 1,080 |
| Nov 25 | 3,970 | 6,200 | 1,160 | 1,100 |
| Dec 25 | 4,010 | 7,300 | 1,180 | 1,125 |
| Jan 26 | 4,050 | 8,200 | 1,200 | 1,140 |
| Feb 26 | 4,080 | 9,000 | 1,220 | 1,155 |
| Mar 26 | 4,100 | 9,700 | 1,235 | 1,170 |
| Apr 26 | 4,110 | 10,200 | 1,248 | 1,184 |
| May 26 | 4,118 | 11,021 | 1,260 | 1,190 |
THE VERDICT
The signals point in the opposite direction: Cursor remains a real, still-growing category leader with strong reach and entrenchment. The pressure is that momentum has rotated toward Claude Code, which sits one layer below Cursor in the stack.
MOMENTUM
ATC's composite momentum blends growth pace, activity volume, acceleration, and adoption-observed company breadth on a 0-100 scale.
Breakout
Developer activity and discussion accelerated fastest after launch.
Strong
Broad enterprise-adoption signal footprint remains the largest in the set.
Strong
Balanced profile across reach, discussion, and developer activity.
Strong
Visible discussion and developer activity, but much narrower enterprise breadth.
TRAJECTORY
The monthly series shows when Cursor's growth normalized, when Claude Code broke out, and when enterprise-adoption activity inflected.
Monthly companies with newly observed public enterprise-adoption signals.
GitHub Copilot starts from the broadest enterprise-adoption signal base. Cursor and Claude Code inflect upward during 2026, while Windsurf remains much smaller.
A sharp early-2026 inflection for both Cursor and Claude Code suggests adoption moving from individual developers to broader enterprise rollout signals.
Company breadth is outside-in adoption-associated evidence. It is not a contract registry or seat count.
Log scale omits zero and missing points from the plotted line.
Source and methodology| Month | Cursor | Claude Code | GitHub Copilot | Windsurf |
|---|---|---|---|---|
| Jun 24 | 0 | missing | 58 | missing |
| Jul 24 | 0 | missing | 75 | missing |
| Aug 24 | 0 | missing | 82 | missing |
| Sep 24 | 18 | missing | 70 | 4 |
| Oct 24 | 20 | missing | 78 | 5 |
| Nov 24 | 0 | missing | 81 | 5 |
| Dec 24 | 32 | missing | 72 | 6 |
| Jan 25 | 48 | missing | 95 | 8 |
| Feb 25 | 72 | 10 | 120 | 10 |
| Mar 25 | 96 | 25 | 145 | 11 |
| Apr 25 | 118 | 50 | 160 | 14 |
| May 25 | 150 | 85 | 180 | 15 |
| Jun 25 | 175 | 140 | 290 | 18 |
| Jul 25 | 205 | 190 | 340 | 21 |
| Aug 25 | 230 | 240 | 380 | 18 |
| Sep 25 | 250 | 280 | 430 | 22 |
| Oct 25 | 240 | 310 | 470 | 13 |
| Nov 25 | 260 | 305 | 495 | 20 |
| Dec 25 | 278 | 315 | 515 | 15 |
| Jan 26 | 291 | 318 | 522 | 17 |
| Feb 26 | 390 | 520 | 610 | 24 |
| Mar 26 | 520 | 760 | 820 | 26 |
| Apr 26 | 650 | 840 | 780 | 18 |
| May 26 | 780 | 980 | 900 | 12 |
POSITION
Cursor's strength is footprint. Momentum and footprint are related but distinct signals.
Current velocity and acceleration.
Accumulated observed activity or company breadth.
| Product | cumulative companies with enterprise-adoption signals |
|---|---|
| GitHub Copilot | 2,820 |
| Claude Code | 2,379 |
| Cursor | 2,025 |
| Windsurf | 92 |
| Product | cumulative developer-activity signal |
|---|---|
| Claude Code | 11,021 |
| Cursor | 4,118 |
| Windsurf | 1,250 |
| GitHub Copilot | 1,207 |
Cumulative footprint does not imply active monthly users, contracts, seat counts, or internal deployment records.
THESIS
Cursor shows 2,025 companies with enterprise-adoption signals and 2,468 companies carrying a broader Cursor-associated signal, including Halliburton, Under Armour, MITRE, and Andreessen Horowitz.
Discussion, usage, and adoption-associated signals move together rather than depending on one isolated channel.
Developer activity rose 112 percent over 24 months and remains visible. Enterprise-adoption activity inflected upward in early 2026.
Cursor's developer-activity growth is real, but it is slower than the fastest-moving peer in this recorded set.
The product gaining share fastest sits one layer below Cursor. Category consolidation toward model labs is the structural threat.
The developer corpus surfaces pricing backlash and questions about differentiation when the underlying models are similar.
VOICE OF DEVELOPER
646 analyzed public comments over the trailing year. Approximately 80 percent Hacker News and 20 percent Reddit over the trailing year.
Observed comments were grouped into themes by ATC and summarized by the model. Direct quotations are not included in this public version.
Developers describe moving from Cursor to Claude Code or Codex after a step-change in output quality.
The moat question appears repeatedly: what does Cursor offer over a code editor when the underlying models are the same?
Billing and packaging complaints appear as a recurring friction point that could increase churn.
Counterweight themes describe Cursor embedded in production workflows and in compliance-sensitive engineering contexts.
EVIDENCE
A sample of companies where ATC observed recent public signals associated with Cursor. These are outside-in adoption indicators, not confirmed contracts, seat counts, or internal deployment records.
TOOL TRACE
ATC supplies structured research data and derived metrics. The AI model produces the narrative and presentation.
The user asked the exact prompt shown on this page.
ATC MCP returned structured product momentum, chart data, company evidence, developer voice themes, and product relationships.
The model synthesized a market-analysis narrative from those structured responses.
A human review removed internal evidence identifiers and withheld raw responses that were not marked public.
The public page adapts the report into responsive HTML, static SVG charts, accessible summaries, and data tables.
Is Cursor overhyped? Compare Cursor, Claude Code, GitHub Copilot, and Windsurf across developer activity, enterprise-adoption signals, product discussion, adoption-observed company breadth, and time-series momentum. Return a verdict, supporting and conflicting evidence, chart-ready trajectories, developer discussion themes, company evidence, and methodology limits.analyze_product_momentumReturned comparative momentum scores, monthly trajectories, product rankings, and reach-versus-momentum inputs.
{
"products": [
"cursor",
"claude-code",
"github-copilot",
"windsurf"
],
"category": "ai-coding-tools",
"start_month": "2024-06",
"end_month": "2026-05",
"metrics": [
"developer_activity",
"enterprise_adoption_activity",
"product_discussion",
"company_breadth"
]
}get_product_comment_evidenceGrouped 646 public developer comments into approved themes without publishing direct quotations.
{
"products": [
"cursor",
"claude-code",
"github-copilot",
"windsurf"
],
"window": "trailing_12_months",
"sources": [
"hacker_news",
"reddit"
],
"include_direct_quotes": false
}get_product_usage_evidenceReturned approved company evidence records with public summaries and source categories.
{
"product": "cursor",
"window": "recent",
"limit": 12,
"public_fields_only": true
}get_product_relationshipsReturned relationship context used by the model to frame substitution and disintermediation risk.
{
"product": "cursor",
"related_products": [
"claude-code",
"github-copilot",
"windsurf"
],
"relationship_types": [
"substitution",
"complement",
"competitive_overlap"
]
}DISCUSSION IN CONTEXT
Discussion activity shows what people are talking about. ATC also examines developer activity, enterprise-adoption signals, adoption-observed company breadth, product relationships, time-series trajectory, and supporting or conflicting evidence.
Mention-frequency and search-frequency views can be useful. ATC is complementary: it is built to connect attention to broader software-market evidence.
METHODOLOGY
Public developer, open-source, repository, community, or related activity as defined in the approved methodology. It is not equivalent to active seats.
Public role, work-description, initiative, or related evidence associated with a product and company. It is not a contract registry.
Unique companies with qualifying adoption-associated evidence under the approved methodology.
Public discussion activity associated with the product, after approved deduplication and resolution.
A derived measure combining approved growth, volume, acceleration, and company breadth components.
The final explanatory prose, thesis, and report structure produced by the AI model from structured responses.
ATC is an outside-in research system. It should not be treated as a contract registry, CMDB, procurement database, product telemetry system, or implementation audit.
A market conclusion is a research thesis, not ground truth.
OTHER QUESTIONS
These prompt cards are static examples. They do not execute on this page.
Compare Snowflake, Databricks, ClickHouse, DuckDB, and Dremio.
Which observability platforms are accelerating?
Compare Vercel, Cloudflare, Netlify, and Fly.io.
Which data-infrastructure products show developer-led versus enterprise-led momentum?
What are developers saying about a product's moat, pricing, and workflow entrenchment?
Which companies show recent adoption-associated signals for a product?
RUN YOUR OWN ANALYSIS
Connect ATC to Claude or ChatGPT to run new comparisons, investigate products, inspect company evidence, and explore software ecosystems.
https://mcp.atc-analytics.com/mcpChatGPT custom connector availability depends on account and workspace settings. See the MCP docs for the current setup path.
ChatGPT connector notesSECURITY, PRICING, LIMITS
SECURITY
PRICING
LIMITATIONS
WHY WE BUILT THIS
Our team came from technology-ecosystem research for institutional investors. The work involved combining weak or conflicting signals into a defensible view of how software categories and companies were changing.
Language models are good at writing an analysis once they have the right context. The difficult part is supplying structured, current evidence about products, companies, developer activity, adoption, and change over time.
We built ATC to expose that research layer as tools a model can call.
FEEDBACK
We would particularly value feedback on whether the signals and methodology are sufficient to audit the conclusion, which software-market questions are most useful, and whether the MCP tool boundaries expose the right level of detail.