Jenkins has been around since 2004, which in software makes it older than most of the tools that replaced it. It is an open-source automation server written in Java, MIT-licensed, and it still sits at the heart of CI for a very large share of enterprise engineering organizations. The pitch has never changed: install it once, connect it to everything through 2,300+ plugins, and build a pipeline for literally any toolchain - including the obscure legacy SDK that no hosted CI will ever support. The trade has not changed either: you get unlimited everything for $0 in software, and you pay for it with your own operating labor.
Pricing: the real math
The software is free. Permanently. No seat limits, no feature gates, no credit meters, no paid tier that unlocks anything - the MIT license means you can run the full thing on a Raspberry Pi or a Fortune 500 cluster and owe nobody anything.
The costs you actually carry:
- Controller: a 4 vCPU / 8 GB VPS runs most mid-size setups - roughly $10-40/mo from Hetzner, DigitalOcean, or $0 on hardware you already own.
- Build agents: ephemeral Docker agents on the controller, cloud VMs billed per minute only while jobs run, or your own machines at zero marginal cost.
- Maintenance labor: plugin updates, security advisories (published roughly weekly), LTS upgrades every 12 weeks. Budget 2-5 hours a month for a healthy instance; a neglected one with an old plugin tree can eat a full day when something finally breaks.
- Commercial support: CloudBees CI, the commercial Jenkins distribution with high availability, governance and SLAs, is quote-only - there is no public price list. Independent consultancies sell Jenkins support contracts the same way.
A realistic all-in for a 10-developer team: $30-100/mo of infrastructure plus the labor. Compare that against GitHub Actions or CircleCI at $15-30/user/mo plus metered minutes, and the software math favors Jenkins badly. The labor math is where the argument gets interesting - and where the business lives.
The money angle: sell the work, not the tool
Nobody resells free software. You sell what teams need around it, and two very different markets exist side by side:
- Migration projects: a large slice of the industry is quietly leaving Jenkins for GitHub Actions, GitLab CI or CircleCI. Converting declarative pipelines, mapping plugin behavior to native features, migrating credentials and runners - a pipeline family of 20-50 jobs runs $2,000-8,000 as a fixed-price project. The freshest demand comes from teams reading any modern CI comparison and realizing their 40-minute Jenkins queue is the bottleneck.
- Managed-Jenkins retainers: the teams that are NOT migrating - on-prem constraints, legacy plugins, or plain institutional inertia - still need the thing patched and green. $1,000-3,000/mo per client to handle upgrades, advisories and pipeline breakage is an easy pitch, because the alternative is a developer doing it badly on Fridays.
- Pipeline rescues: multi-hour builds, flaky agents, disk-full controllers, zombie jobs. Flat-fee audits at $1,000-2,500 with a short report and a fix list convert well because the pain is loud and specific.
- Shared libraries and custom plugins: companies with genuinely unusual toolchains pay $3,000-10,000 for a well-built Groovy shared library or plugin that their internal team cannot maintain.
The honest framing: migration work is one-time revenue, and the retainers come from the teams that will never migrate. Both markets are large. If you already do DevOps consulting, Jenkins work is the highest-margin line item in it - the software is free, so every dollar a client pays is for your judgment, not a license markup.
Where it falls short
Be clear-eyed before picking it:
- Plugin decay is the structural risk. 2,300+ plugins, many maintained by one person. Upgrades conflict, projects get abandoned, and version pinning becomes a survival skill. Every Jenkins operator has a story about a minor plugin update taking the controller down.
- The UI has not aged well. Blue Ocean was the promised redesign, went quiet years ago, and the classic interface is functional at best. Nobody chooses Jenkins for the experience.
- You own the security surface. An exposed or unpatched Jenkins controller is a classic breach vector, and plugin CVEs need weekly attention. Hosted CI competitors absorb this; you cannot opt out of it.
- Performance is DIY. Build queues, agent autoscaling, cache strategy - hosted CI hands you all of this configured. On Jenkins it is configuration work, and the default is a slow queue everyone complains about.
- Scripted Groovy ages badly. The 400-line pipeline file nobody dares touch is a cliche because it is real. Declarative pipelines plus shared libraries mitigate it, but legacy estates are full of the other thing.
Who should pick it
Pick Jenkins if you need full environmental control - air-gapped networks, on-prem data policies, agents on exotic hardware - or if you already run a large estate that works, because ripping it out rarely pays for itself. Pick a hosted alternative if you are greenfield: GitHub Actions for zero-maintenance simplicity, CircleCI for fast parallel test splitting, GitLab if you want CI bundled with your repos and security scanning. The full trade-offs are laid out in our CI/CD comparison.
And if you are already living with a creaky Jenkins instance: that pain is somebody's business model, and it could be yours.
FAQ
Is Jenkins really free? What is the catch?
Yes, genuinely free under the MIT license - every plugin, unlimited users and jobs, nothing locked. The catch is operational: you provision and patch the controller, manage plugin upgrades and security advisories, and debug the pipeline when an update breaks it. A 10-developer team realistically spends $30-100/mo on infrastructure and a few hours a month on upkeep. Teams that cannot spare that either buy CloudBees CI (quote-only commercial distribution) or pay a consultant - a real, functioning market.
Why do enterprises still run Jenkins in 2026?
Depth, control and freedom. Hundreds of pipelines and shared libraries accumulated over a decade are expensive and risky to migrate. Air-gapped networks, on-prem policies and exotic build hardware fit Jenkins and sit badly with hosted CI. And there is no metered bill that grows with headcount. Whether those reasons still justify the labor is exactly the conversation migration consultants get paid to have.
When does Jenkins stop making sense?
When nobody on the team owns its upkeep. Greenfield teams without a dedicated ops person are better served by GitHub Actions or CircleCI, where the controller, security patching and runner fleet are someone else's job. If your team spends more time maintaining the CI than the software it builds, that is the signal - migrate, or hand the instance to someone who runs them for a living.