~ / AI Research / CI/CD Market Analysis

CI/CD Market Analysis

Comprehensive analysis of the Continuous Integration / Continuous Delivery market — from legacy incumbents (Jenkins, TeamCity) to platform-bundled CI (GitHub Actions, GitLab CI) to dedicated challengers (CircleCI, Buildkite, Harness) to the new wave of AI-powered and niche tools (Trunk, Depot, Blacksmith, Dagger). Revenue numbers, funding data, user counts, pricing models, and strategic analysis for each major player.

The core question: GitHub Actions has commoditized basic CI/CD by bundling it free with source control. Jenkins still holds ~48% market share despite being painful to operate. Where are the real opportunities for new entrants, bootstrappers, and focused products?

Section 1: Market Overview

Total Market Size

Market sizing varies significantly across analyst reports due to differing definitions of what counts as “CI/CD”:

Best estimate: The narrowly-defined CI/CD tooling market is roughly $8–15 billion in 2024, growing at 11–21% CAGR depending on inclusion criteria. The broader DevOps market (which includes CI/CD) was valued at $11.8B in 2024, expected to reach $64.4B by 2033 at 20.7% CAGR.

Deployment Split

Cloud-based CI/CD captured 63–67% of total revenue in 2023/2024. Self-hosted remains significant in enterprises with strict compliance requirements (finance, healthcare, government, defense). Hybrid models are growing — use a cloud orchestrator (like GitHub Actions) but run on self-hosted runners.

Market Evolution Timeline

2004–2014: The Jenkins Era

Hudson/Jenkins dominates. CI is a self-hosted Java application you babysit. Everyone writes XML pipeline configs. Painful but no alternatives.

2011–2017: Cloud CI Emerges

Travis CI (2011) pioneers free CI for open source on GitHub. CircleCI (2011), Codeship, Semaphore follow. SaaS CI becomes the norm for startups. Travis CI becomes ubiquitous in open-source — the .travis.yml badge is everywhere.

2017–2019: Platform Consolidation

GitLab ships integrated CI (2015, matures by 2017). Bitbucket adds Pipelines. Azure DevOps evolves from TFS/VSTS. The “all-in-one DevOps platform” thesis takes hold. Travis CI acquired by Idera (Jan 2019) and begins declining.

2019–2022: GitHub Actions Changes Everything

GitHub Actions launches (Nov 2019) with generous free tier. Being bundled with the world’s largest code host gives it instant distribution. Dedicated CI companies start losing open-source and SMB customers. CircleCI, Travis CI, and others feel the squeeze.

2022–2024: Specialization & GitOps

Argo CD and Flux CD graduate in CNCF. GitOps becomes standard for Kubernetes. Monorepo tools (Nx, Turborepo) grow. Supply chain security (SLSA, Sigstore, SBOM) becomes mandatory. Platform engineering emerges as a discipline.

2024–2026: AI Wave & Niche Unbundling

AI enters CI/CD: predictive test selection, auto-fix, pipeline generation. GitHub Actions raises prices for self-hosted runners (2026). Third-party runner providers (Depot, Blacksmith, Namespace) emerge. The market unbundles into specialized layers: build acceleration, test analytics, merge queues, ephemeral environments.

Section 2: Major Players

Big Platform CI/CD (Bundled)

GitHub Actions (Microsoft)

GitLab CI/CD (GTLB, Public)

Bitbucket Pipelines (Atlassian)

Azure DevOps Pipelines (Microsoft)

Dedicated CI/CD Platforms

CircleCI

Travis CI (Idera)

Jenkins (Open Source) / CloudBees (Enterprise)

TeamCity (JetBrains)

Buildkite

Semaphore CI

Drone CI / Gitness / Harness Open Source

Woodpecker CI

Concourse CI

Tekton

Dagger (Solomon Hykes / Docker Founder)

Earthly

Depot

Blacksmith

Namespace

RWX / Mint CI

Nx Cloud (Nrwl)

Turborepo (Vercel)

Cloud/Infrastructure CI

AWS CodeBuild / CodePipeline

Google Cloud Build

Harness

Codefresh (acquired by Octopus Deploy)

Octopus Deploy

Spacelift

env0

Argo CD (Open Source)

Flux CD (Open Source)

Spinnaker (Netflix)

Mobile CI/CD

Bitrise

Codemagic

Appcircle

Fastlane (Google / Open Source)

Microsoft App Center

Security/Compliance CI

Snyk

SonarSource (SonarQube / SonarCloud)

Checkmarx

Veracode

AI-Powered CI/CD (New Wave)

Trunk.io

Aviator

BuildPulse

Currents

Launchable (acquired by CloudBees)

Section 3: The Jenkins Story

Jenkins by the Numbers

• Market share: 44–48% of CI/CD. Still #1 by a massive margin.
• 1,800+ plugins
• Millions of installations worldwide
• Created in 2004 by Kohsuke Kawaguchi as “Hudson” at Sun Microsystems
• Forked to Jenkins in 2011 after Oracle acquired Sun

Why Jenkins Persists Despite Being Painful

1. Massive installed base: Decades of custom pipelines, scripts, integrations. Migration cost is enormous.
2. Plugin for everything: 1,800+ plugins means Jenkins can integrate with any tool, any cloud, any language.
3. Self-hosted = total control: No vendor lock-in, no data leaving your network. Critical for regulated industries.
4. “It works, don’t touch it” mentality: Jenkins pipelines that have been running for years are treated as untouchable infrastructure.
5. Free (as in beer): Open-source, MIT license. No per-user or per-minute fees.
6. Tribal knowledge: Every DevOps engineer knows Jenkins. Hiring is easier. Documentation is abundant.

The Jenkins Pain Points

• Groovy-based Jenkinsfiles are notoriously difficult to debug and test
• Plugin hell: Plugins conflict, break on updates, have security vulnerabilities, and are often abandoned
• Scaling is hard: Jenkins master becomes a bottleneck. Kubernetes agents help but add complexity
• UI is ancient: The Jenkins UI looks like 2008. Blue Ocean was an attempt to modernize but was abandoned
• Maintenance burden: Someone has to keep the Jenkins server patched, backed up, and running. This is a full-time job at scale
• Total Cost of Ownership: Jenkins is free but hosting, maintaining, and operating it is not — and costs are unpredictable

CloudBees: Monetizing Jenkins Pain

CloudBees has built a $150M+ ARR business by selling managed, enterprise-grade Jenkins. $553M in funding, unicorn valuation, 500+ enterprise customers. They’ve expanded beyond Jenkins into a broader software delivery platform (feature flags, release management, analytics) but Jenkins remains the core wedge. The acquisition of Launchable (AI test selection by Jenkins’s own creator) shows CloudBees investing in making Jenkins pipelines smarter, not just more manageable.

The “Jenkins Replacement” Market

Every CI company pitches “migrate from Jenkins” as a use case. But Jenkins replacement is incredibly hard to sell because:

• Migration requires rewriting every pipeline, often hundreds or thousands of them
• The people who maintain Jenkins are the same people who would lead the migration — and they’re already overworked
• Budget holders see Jenkins as “free” and resist paying for a replacement
• Jenkins custom plugins have no equivalent in other platforms
• The risk of migration (breaking CI = breaking all deployments) is high

The realistic Jenkins exit path: New projects start on GitHub Actions or GitLab CI. Old Jenkins pipelines are gradually migrated or left to rot. Full replacement takes 3–5 years per organization. This is why Jenkins market share declines slowly, not quickly.

Section 4: GitHub Actions Impact

How GitHub Actions Changed CI/CD Forever

1. Killed the standalone free tier: When the world’s largest code host gives you 2,000 free CI minutes/month, why pay for Travis CI or CircleCI’s free tier? Travis CI died within 2 years of GHA launch.
2. Made CI a feature, not a product: CI/CD went from being a separate purchasing decision to being bundled with your code host. Like how browsers killed standalone RSS readers.
3. YAML as lingua franca: GitHub Actions’ YAML workflow syntax became the de facto standard. Love it or hate it, everyone knows it.
4. Marketplace created ecosystem lock-in: 20,000+ Actions in the marketplace. Using them creates switching costs to other CI platforms.
5. Forced everyone to have a free tier: Every CI platform now must offer generous free usage or risk irrelevance.

GitHub Actions Usage Stats

• 5M+ workflows per day, up 40% YoY
• 62% of developers use GHA for personal projects (JetBrains 2025 survey)
• 41% use GHA in organizations
• Docker/Kubernetes usage in Actions workflows up 25% (~1M containerized workflows)
• Deployment times cut 30% on average

Who Lost Customers to GHA?

GHA Limitations That Create Opportunities

• Slow runners: Standard GHA runners are 2-core machines. Large builds are painfully slow. This created the Depot/Blacksmith/BuildJet market.
• No native merge queue (until recently): GitHub shipped a basic merge queue, but Trunk and Aviator offer significantly more sophisticated alternatives.
• No test analytics: GHA doesn’t tell you which tests are flaky, slow, or redundant. This creates the BuildPulse/Currents/Trunk analytics market.
• No remote caching: Build caching in GHA is basic. Nx Cloud, Turborepo, and Depot offer much better caching.
• Security concerns: Third-party Actions run opaque code with access to secrets. No built-in supply chain verification.
• Log viewer crashes: The GHA web UI is buggy for large builds. Browser crashes on long logs.
• YAML complexity: Complex workflows in YAML become unreadable. No good testing/debugging story for workflow files.

GitHub Actions 2026 Pricing Changes

Starting January 2026, GitHub reduced hosted runner prices by up to 39%. But starting March 2026, they introduced a $0.002/minute platform fee for self-hosted runners. The Actions control plane is no longer free. This is significant because many large companies were using GitHub’s orchestration while running on their own hardware, paying GitHub nothing for CI execution.

96% of customers see no bill change. But this signals GitHub will increasingly monetize CI/CD as a revenue center, not just a GitHub retention tool.

Section 5: Emerging Trends

AI in CI/CD

• Predictive test selection: ML models analyze code changes + historical data to determine which tests to run. Launchable (now CloudBees), Trunk, and others. Companies report 50%+ reduction in test cycle time.
• Auto-fix of CI failures: AI agents that can diagnose and fix broken builds. Still early but emerging in GitHub Copilot, Harness AIDA, and specialized tools.
• Pipeline generation: AI generating CI/CD pipeline configs from natural language descriptions. GenAI streamlining boilerplate code, CI/CD configs, and IaC files.
• Flaky test detection: ML identifying tests that fail randomly (not due to code changes). BuildPulse, Trunk, Currents, Launchable all offer this.
• Intelligent test parallelization: Beyond round-robin splitting — ML-optimized distribution of tests across CI machines based on historical duration and flakiness.

GitOps Adoption

• Argo CD used by 60% of Kubernetes clusters in CNCF survey
• 97% of Argo CD users run it in production
• 100% of CNCF survey respondents plan to adopt GitOps within 2 years
• Platform engineers represent 37% of Argo CD users
• Argo CD is winning decisively over Flux CD in adoption

Platform Engineering & Internal Developer Platforms

• Gartner predicts 80% of large software orgs will have platform engineering teams by 2026 (up from 45% in 2022)
• 94% of surveyed enterprises see AI as essential to platform success
• 75% of developers lose 6+ hours weekly to tool fragmentation
• Backstage (Spotify), Port, OpsLevel, Humanitec, Cortex are the leading IDP tools
• CI/CD is increasingly consumed as a service within the platform rather than a standalone tool

Monorepo Tooling

• Nx: 1.5M+ developers, $24.6M funding, expanding beyond JS to Java/C#/.NET
• Turborepo (Vercel): Free remote cache, tight Vercel integration, JS/TS focused
• Bazel (Google): Powerful but adoption limited to large companies due to steep learning curve and migration difficulty
• Buck2 (Meta): Open-sourced 2023 but still immature outside Meta. Limited docs, small community
• Pants: Python-focused alternative to Bazel. Niche but loyal users
• Key insight: Declarative build systems (Bazel, Buck2) won’t get traction in open source because they’re painful for small projects. Nx and Turborepo win because they’re approachable.

Supply Chain Security

• SLSA 1.0 (Supply-chain Levels for Software Artifacts) is stabilized
• Sigstore (cosign) enables keyless code signing, widely adopted
• SBOM (Software Bill of Materials) now legally required by EU Cyber Resilience Act (CRA)
• US federal agencies required to comply with Secure Development Framework (SSDF)
• Most organizations stall at SLSA Level 2 — tooling is mature but organizational adoption lags
• Large buyers increasingly demand provenance (proof of how software was built) alongside SBOMs

Merge Queues

• GitHub shipped a native merge queue (basic)
• Trunk, Aviator, Mergify offer more sophisticated alternatives
• Critical for monorepos where main branch stability is essential
• Trunk reports up to 90% CI cost reduction from intelligent queuing

Ephemeral Environments

• Gartner 2024: organizations using ephemeral environments increasing rapidly
• Uber, Lyft, DoorDash moving beyond shared staging to isolated preview environments
• Platforms: Bunnyshell, Vercel (preview deployments), Signadot, Humanitec
• Each PR gets its own environment — eliminates staging bottlenecks

CI/CD for AI/ML Pipelines (MLOps)

• MLOps market: $1.58B (2024), projected $2.33B (2025) at 35.5% CAGR
• 55%+ of AI teams use automated CI/CD for model deployment
• 85% of ML models never make it to production — the pipeline gap is enormous
• Traditional CI/CD tools don’t handle data versioning, model training, experiment tracking
• Specialized tools: MLflow, Kubeflow, DVC, Weights & Biases, Neptune.ai

Feature Flags Integration

• LaunchDarkly serves 4B+ feature flags daily
• Feature flags decouple deployment from release — deploy code to production behind a flag, enable gradually
• Harness, GitLab, and others now bundle feature flags into their CI/CD platforms
• Increasingly considered a core part of CD, not a separate tool

Remote Build Caching / Execution

• Turborepo Remote Cache (free on Vercel) — reduces CI times up to 85%
• Nx Cloud — distributed task execution + remote caching across CI machines
• Gradle Build Cache — built into Gradle ecosystem
• Depot — persistent Docker layer caching across builds
• Bazel Remote Execution (BES/RBE) — Google-scale but complex to set up

Section 6: Pricing Models

Pricing Model Comparison

The Race to Zero

GitHub Actions’ generous free tier forced every CI company to offer free usage. But free tiers are expensive to operate. The economics:

• GitHub can subsidize CI because it monetizes through Copilot ($19–$39/user/mo) and Enterprise subscriptions
• GitLab bundles CI into platform pricing — CI is a feature, not the product
• Standalone CI companies (CircleCI, Semaphore) can’t afford the same generosity
• This is why focused “acceleration” tools (Depot, Blacksmith) work — they charge on top of GHA, not instead of it

Self-Hosted vs Cloud Cost Comparison

At scale (50+ developers, thousands of builds/day):

• Self-hosted Jenkins: 0.5–2 FTE for maintenance ($75K–$300K/yr) + server costs ($5K–$50K/yr) + plugin management overhead. “Free” but TCO is $100K–$400K/yr.
• GitHub Actions (hosted): $10K–$100K/yr depending on usage. No maintenance overhead. But slow runners may cost in developer time.
• Buildkite (hybrid): $15/user/mo + your own compute. Predictable orchestration cost, flexible compute.
• GitLab (SaaS): $29–$99/user/mo. All-inclusive but expensive per-seat at scale (100 devs at $99 = $119K/yr).

Section 7: The Bootstrapper / DHH Angle

Open-Source Self-Hosted CI/CD Options

What Gaps Exist for Bootstrappers?

The CI/CD market looks saturated but there are specific underserved niches where small, focused products can win:

1. “Picks and Shovels” Plays (Proven Market)

Test Analytics / Flaky Test Detection

BuildPulse, Currents, and Trunk Analytics exist but the market is early. Every team with 1,000+ tests has flaky test pain. Standalone test analytics that works with any CI is a viable bootstrapped business. Example: BuildPulse appears to be a small, focused team doing this.

Build Caching as a Service

Depot proved this works ($5M revenue on $4.6M funding). Docker layer caching, build artifact caching, and remote execution are all valuable. Turborepo’s free cache helps JS teams but other ecosystems (Go, Rust, Python) lack good caching solutions.

CI Cost Analytics / Optimization

Teams don’t know what their CI costs or where the time goes. A tool that analyzes CI spend across GitHub Actions, GitLab, etc. and recommends optimizations could be valuable. Trunk does some of this but it’s bundled with their larger platform.

Pipeline Visualization / Debugging

GitHub Actions’ log viewer crashes browsers. Complex YAML pipelines are unreadable. A better visualization tool for CI pipelines (across providers) is an underserved need.

Merge Queues

Trunk ($28.5M funded) and Aviator ($2.4M funded) prove demand. But GitHub’s native merge queue is catching up. The window may be closing unless you differentiate significantly.

2. Vertical-Specific CI/CD

Mobile CI/CD

App Center is retiring (Mar 2025). Bitrise is $95M funded and Codemagic is niche. There’s room for a focused, affordable mobile CI that does iOS + Android well. The macOS runner requirement creates a natural moat (expensive infrastructure).

Embedded / IoT CI

Firmware testing, hardware-in-the-loop CI, and cross-compilation pipelines are poorly served by general-purpose CI. Niche but high-value customers.

Data Pipeline CI

Testing dbt models, Airflow DAGs, Spark jobs. The MLOps space ($1.58B) needs better CI but traditional CI tools don’t understand data dependencies.

Infrastructure-as-Code CI

Spacelift ($82.3M funded) and env0 ($42M funded) prove this is a real market. But they’re VC-funded and complex. A simpler, cheaper IaC CI for small teams could work.

Game Development CI

Unity/Unreal builds are massive and have unique requirements (asset pipelines, shader compilation, platform-specific packaging). GameCI exists as open-source GitHub Actions but no commercial product dominates.

3. The Octopus Deploy Model (Bootstrapper Gold Standard)

Octopus Deploy is the proof case: Paul Stovell bootstrapped a CD-focused tool for .NET/Windows for 11 years, grew to $80M+ ARR, then took VC to accelerate. Key lessons:

• Picked a neglected niche (CD for .NET/Windows when everyone focused on Linux/Docker)
• Built a sustainable business without venture capital for over a decade
• Took funding only when ready to acquire (bought Codefresh for $43M) and expand
• Now at $882M valuation with 270+ employees and 4,000+ customers

4. Solo-Founder CI/CD Products That Could Work

• CI Status Dashboard: Aggregate CI status from GHA, GitLab, Jenkins into one view for engineering managers. Charge $20–$50/team/month.
• CI Cost Reporter: Connect to GitHub/GitLab billing APIs, show cost per repo/team/workflow, suggest optimizations. Every CFO wants this visibility.
• Self-Hosted Runner Manager: Tool to manage a fleet of self-hosted GitHub Actions runners (auto-scaling, monitoring, updates). Currently requires custom scripting.
• Pipeline Linter / Validator: Validate and optimize GitHub Actions / GitLab CI YAML before committing. Catch errors before they waste CI minutes.
• Flaky Test Tracker: Simple SaaS that ingests test results from any CI, identifies flaky tests, and provides actionable dashboards. BuildPulse does this but the market can support multiple players.
• CI/CD for Hugo/Jekyll/Static Sites: Ultra-simple CI for static site generators. Most are over-engineered for this use case.
• Deployment Notifications: Rich Slack/Teams/Discord notifications about deployments (who deployed what, when, diffs, rollback links). Surprisingly few good solutions.

Key Takeaways for Bootstrappers

1. Don’t build another CI platform. GitHub Actions, GitLab CI, and Jenkins have distribution you can’t match. CircleCI raised $315M and is struggling.
2. Build on top of existing CI. The Depot model ($5M revenue, $4.6M funding) is ideal: make GitHub Actions faster/better/cheaper without replacing it. One-line migration.
3. Sell to the pain, not the feature. “Your CI is slow” ($) / “Your tests are flaky” ($$) / “You don’t know what CI costs” ($$$) are better pitches than “our CI is better.”
4. Target a neglected ecosystem. Mobile (App Center retiring), embedded/IoT, game dev, data pipelines, IaC — all underserved by general-purpose CI.
5. Self-hosted is a feature, not a bug. Enterprises in regulated industries will pay premium for tools they can run behind their firewall.
6. Follow the Octopus Deploy playbook. Pick a niche, bootstrap for years, build a real business, take funding only when you’re ready to acquire/expand.

Revenue Comparison Table

Security-Adjacent Revenue Table

IaC CI/CD Revenue Table

Research compiled February 2026. All revenue, funding, and valuation figures sourced from company announcements, SEC filings, Crunchbase, PitchBook, GetLatka, Tracxn, and news reports. Figures may reflect different reporting periods.