$92 million in 2023. $7.4 billion in 2025. One of the fastest category formations in software history.
The vibecoding market grew 798% in two years. What was a niche in early 2023 — AI-powered code generation used by a small community of early-adopter developers — had, by early 2025, become a $7.4 billion category absorbing developer and enterprise spend at an extraordinary rate. Collins Dictionary named "vibe coding" its Word of the Year for 2025. Andrej Karpathy coined the term in February 2025 in a post that reached 5.1 million views. Within twelve months, 92% of US developers were using AI coding tools daily. The adoption curve is not done: only 0.6% of the 4.5 billion global working-age population has tried an AI coding tool.
This report covers the full vibecoding market — from professional IDE tools to consumer app builders, from CLI-native coding agents to platform-embedded features — across 21 pages and 58 sources. It sizes the market from the bottom up, maps the competitive landscape across four player archetypes, documents enterprise adoption data from named deployments, and models the addressable base expansion through 2030.
What the analysis covers
The report examines the market across four competitive archetypes — Professional IDE Tools (Cursor, GitHub Copilot), Consumer App Builders (Lovable, Bolt), CLI/Terminal Agents (Claude Code, Devin, Codex), and Platform/Embedded players (Amazon Q, Gemini Code Assist) — and two revenue segments: B2B (68% of revenue at $5.0B) and B2C (32% at $2.4B). Company-level analysis covers ARR trajectories, market share, ARPU structure, moat durability assessments, and go-to-market positioning. Market sizing covers 2023–2025 actuals with base, bull, and bear projections through 2030. Five white spaces are identified and sized for the post-2026 opportunity landscape.
Three players, 74% of market — with volume, not pricing, as the growth engine
The vibecoding market is already highly concentrated. Claude Code holds approximately $2.5B ARR and 34% market share; Cursor follows at approximately $2.0B and 27%; GitHub Copilot accounts for approximately $1.0B at 14% share. The top three collectively hold 74% of the market. OpenAI's acquisition of Windsurf for approximately $3 billion in February 2026 signals what comes next: foundation model providers vertically integrating into the tooling layer, compressing the share available to independent players without structural distribution or proprietary model advantages.
Despite this concentration, the market is forecast to reach $32.5B by 2030 at approximately 34% CAGR. Critically, ARPU is projected flat from 2026 onward — $300 per year B2C, $368 per seat B2B. The growth engine is volume, not price. The evidence: only 0.6% of the 4.5 billion global working-age population has tried an AI coding tool. If non-developer adoption reaches even a fraction of that addressable base, the TAM expansion dwarfs any pricing improvement. This is a user acquisition story, not a monetisation story.
Enterprise adoption is real — and the productivity data is now documented
Enterprise productivity gains from vibecoding tools are no longer anecdotal. Documented deployments at Walmart, Klarna, Booking.com, and Accenture show productivity improvements of 31–55%, with the variation reflecting differences in deployment model, human review practices, and the proportion of development work eligible for AI acceleration. 87% of Fortune 500 companies now use AI coding tools, and procurement is formalising: the shift from individual subscriptions at $10–20 per month to negotiated enterprise contracts at $39–500+ per seat is already underway, with Cursor reporting approximately 60% of its $2B ARR from enterprise contracts.
But trust in AI-generated code has fallen from 77% to 60% in one year. 45% of AI-generated code contains OWASP Top-10 vulnerabilities. Over 70% of AI-generated Java code fails security checks. The enterprise market in 2026 is not asking whether to adopt — it is asking how to deploy at scale without creating a security liability. The organisations that have built governance frameworks are pulling ahead; those running ad-hoc AI tools are accumulating risk in proportion to their adoption velocity.
Five white spaces remain materially underserved
Despite the category's scale, five segments remain largely unaddressed by current players. Regulated-industry builders — tools purpose-built for financial services, healthcare, and government, where HIPAA, SOX, and FedRAMP requirements apply by default — represent the largest opportunity. No major horizontal platform has committed to a vertical. A coding agent that generates compliant code by default commands an estimated 5–10x pricing premium over horizontal alternatives, in industries with documented $150 billion annual IT spend. Vertical-specific agents represent a second distinct opportunity, with high-compliance sectors willing to pay for domain context that horizontal tools cannot provide.
The remaining three white spaces are AI code assurance (post-generation security scanning purpose-built for AI code vulnerability patterns — distinct from traditional SAST/DAST tools), mobile-first builders (Rork is the only dedicated player in a segment anchored by $90 billion in annual App Store revenue), and legacy code migration (modernising the billions of lines of COBOL, Fortran, and aged Java that enterprises cannot easily retire). Each has documented willingness to pay and weak incumbent positioning from the current leaders.
798% in two years. 0.6% of the addressable population reached. The vibecoding market has already demonstrated what its growth ceiling is not — the question is which players will be standing when non-developer adoption reaches scale.
Why this matters
For technology investors and VCs, the concentration data and white space analysis define where defensible returns are available in 2026–2030. The top three players — Claude Code, Cursor, GitHub Copilot — have established moats that compound with enterprise data and distribution. The white spaces are where first-mover compliance certification or domain-specific training data creates procurement lock-in that horizontal tools cannot replicate quickly. Flat ARPU means growth comes from volume and new user segments — which favours players with the widest addressable bases and the lowest barriers to non-developer onboarding.
For enterprise technology leaders, the documented productivity range of 31–55% provides the business case benchmark, while the OWASP vulnerability data and governance framework analysis defines what a safe deployment actually looks like. The four-archetype competitive map identifies which tools serve which buyers at which price points — enabling tool selection decisions that will hold through the consolidation cycle now clearly underway.
Access the full report
We are publishing this report at no cost. The full PDF includes the complete market sizing model with base, bull, and bear cases through 2030; detailed company profiles and moat assessments for the eight primary players across all four archetypes; the bottom-up TAM expansion model with the 0.6% penetration analysis; productivity benchmark data from four named enterprise deployments; and the five white space analyses with addressable market estimates. You are free to use and reference the findings with attribution to the Autonomous Research Group.
If you'd like to discuss the analysis, request the underlying data in Excel format, or commission a customised version for your sector or portfolio — get in touch at team@something-better.org.