2026 Cost model · Every input shown

What does platform engineering actually cost?

For a 100-engineer organisation with a 6 to 10-engineer platform team, our model puts it at $1.9M to $2.77M per year all-in, or $19k to $28k per developer served. Most of it is salary. Every input is shown and adjustable below.

Total annual cost

$2,770,000

$27,700 per product engineer served

Quick preset:

20Organisation size1000

Platform salaries$1,900,000(69%)

Tooling$500,000(18%)

Cloud / infra$85,000(3%)

Overhead$285,000(10%)

At 100 product engineers, you are looking at a platform team of 10, tooling spend of $500k, IDP infrastructure of $85k, and $285k of overhead. All-in total: $2.77M per year, or $28k per product engineer served.

Platform-to-product ratio

1 : 8-12

Roughly 8-11% of the engineering org. Our planning band, not a survey finding.

Cost per dev served / year

$19k-$28k

Our model at 100 engineers, 6-10 platform engineers. By ratio on /cost-per-developer.

Payback period

Not published

No study gives one. Our planning view and a worked one-pager on /roi.

02 Breakdown

Where the money actually goes

of spend

69-79%

People

Salary and overhead is always the biggest line item. Five engineers at our senior loaded rate is $1.35M a year with overhead.

CI/CD, observability, secrets, IaC orchestration, service catalogue. Generic budget bands. No vendor pricing.

The platform itself runs on cloud compute. Our managed-cloud formula gives $61k to $155k a year from 30 to 300 engineers.

On-call premiums, training, internal advocacy, migration tax. Easy to forget, always there.

Drill in

03 Justification

Is this investment worth it?

The canonical frame is one hour saved per product engineer per day. At 100 engineers, $100 loaded hourly rate, and 220 working days, that is $2.2M per year of recovered capacity, set against $1.9M to $2.77M of platform cost in our 100-engineer model. The case only holds if most of that hour is real and attributable to the platform, so measure it rather than assume it.

DORA's 2024 report found that using an internal developer platform improves individual productivity, team performance and organisational performance, but can also lead to decreased change stability and throughput. DX reports its Core 4 approach implemented at more than 300 companies, with a 3 to 12 percent increase in engineering efficiency.

The honest caveat: under 30 engineers, the math usually does not work. See when to invest.

Build the CFO one-pager Run your numbers

Payback curve (our planning view)

Pure costMonths 0-6
Team hired, no value delivered yet.
First golden paths shipMonths 6-12
Value accrues but does not offset cost.
Critical mass of adoptionMonths 12-18
Value can approach cost if adoption holds.
ROI inflectionMonths 18-24
Value can overtake cost; measure, do not assume.

04 Scaling

How team size evolves with the org

These are our own planning bands, not survey findings. Full treatment, and what the published research does and does not cover, on /team-structure.

Org size	Platform team	Structure	Team people cost
Under 30 eng	0-1	Fractional; senior eng wears the hat part-time	$38k-$76k
30-80 eng	2-4	Dedicated team, single unit	$380k-$760k
80-200 eng	6-12	Team with specialisation (CI/CD, infra, DevEx)	$1.14M-$2.28M
200-500 eng	15-30	Multiple sub-teams plus dedicated manager	$2.85M-$5.7M
500+ eng	30-80	Platform org with its own VP / director	$5.7M-$15M

Frequently asked questions

How much does a platform engineering team cost per year?

For a 100-engineer organisation with a 6 to 10-engineer platform team, our model puts it at $1.9M to $2.77M a year all-in. People (loaded salary plus on-call, training and recruiting overhead) are 69 to 79 percent of that, tooling 18 to 26 percent, and the platform's own cloud 3 to 4 percent. These are outputs of our planning model, not survey data; every input is listed on /cost-for-100-developers and adjustable on /calculator.

How many platform engineers do I need?

Our planning band is a platform-to-product ratio of 1:8 to 1:12, which is roughly 8 to 11 percent of the engineering organisation. That band is our own heuristic rather than a published survey finding: the CNCF Platform Engineering Maturity Model and platformengineering.org publish maturity and adoption data but no staffing ratio, and guidance published elsewhere often runs looser, around 1:15 to 1:25. Minimum viable team size is two to three engineers, below that you have a single point of failure. Specialisation typically emerges around 80 product engineers, and a dedicated manager becomes necessary around 120.

What is the ROI of platform engineering?

No published study gives a payback period for platform teams, so treat any single figure with caution, including ours. The usual frame is time recovered: one hour saved per product engineer per day at a loaded hourly rate of about $100 is $2.2M a year for 100 engineers, against $1.9M to $2.77M of platform cost in our model, so the case rests on the hour being real and attributable. DORA research links mature platforms to higher deployment frequency and lower change-failure rates; DX reports a 3 to 12 percent increase in engineering efficiency at organisations using its Core 4 framework. See /roi for the full model.

When should we start investing in platform engineering?

Under 20 engineers, almost certainly not yet. Use managed services and conventions; a platform team adds overhead you cannot yet absorb. Between 20 and 50 engineers, one fractional engineer can help with pipeline and cloud. From 50 engineers and up, you are paying the cost of not having a platform team whether you acknowledge it or not. Above 80 engineers, failing to invest is the expensive option.

Is it cheaper to build or buy an internal developer platform?

Neither extreme works well. Pure build means carrying a product team forever to maintain internal tooling with no feature parity to commercial options. Pure buy means integration work that rarely shrinks and licence costs that escalate. Most successful platforms at 200+ engineers are hybrids: commercial core for the capabilities vendors do well (CI/CD, observability, service catalogue) and custom glue for the golden paths unique to your stack. See our /build-vs-buy decision framework for the full treatment.

What is the cost per developer served benchmark?

There is no published benchmark for cost per developer served. In our model, at 100 product engineers with a 6 to 10-engineer team, standard tooling and managed cloud, it comes to $19k to $28k per product engineer per year. The figure is driven mostly by the platform-to-product ratio: loosen the ratio and it falls, tighten it and it rises. Our /cost-per-developer page shows the figure at each ratio.

How do platform engineering costs compare to DevOps consulting?

Consulting rates are quoted per engagement and vary too widely by firm, seniority and location for a single figure to be useful, so we do not publish one. The structural difference is that a consulting engagement ends, while a permanent platform team delivers ownership, knowledge retention and iteration. A common pattern is consulting for kickoff or a specific migration, then in-house ownership.

What hidden costs do platform engineering budgets usually miss?

On-call pay, training, recruiting, internal advocacy (roadshows, office hours, scorecard reviews), documentation handover when senior engineers leave, and double-running during migrations. Our model adds 15 percent of salary for the first three; the rest are worth budgeting on top from your own history. Our /hidden-costs page goes through each item.

Adjacent cost guides

Engineering Hiring Cost

Cost of hiring engineers end-to-end.

ML Engineer Salary

Adjacent salary data for ML roles.

Incident Cost

What platform engineering prevents.

CI/CD Cost

Deeper look at one tooling category.

Monitoring Cost

Observability budget reference.

IT Budget Calculator

Parent budgeting framework.

Why you will not find vendor prices on this site

This is an independent cost-framework site. We do not sell platform engineering tools, take sponsorship from IDP vendors, or publish specific vendor pricing. Cost figures are outputs of our own planning model, with every input shown and adjustable on the calculator; public sources (BLS OEWS, DORA, DX, CNCF, vendor pricing pages) are cited where a figure rests on them. Tooling categories are referenced generically so the bands remain useful regardless of who you buy from. Full editorial policy.