Platform engineering budget calculator (2026)
Enter your inputs. Get a modelled annual budget, cost per product engineer served, and a share-of-spend breakdown suitable for a finance review. No vendor names, no demo funnel, no email-gate.
Describe your organisation
Your modelled budget
- Platform salaries$2,850,00069%
- Tooling$750,00018%
- Cloud / infra$102,5002%
- Overhead$427,50010%
How this calculator works
The calculator is a deterministic model with four inputs (org size, loaded cost per platform engineer, target ratio, tooling tier, cloud model) and one optional overhead uplift. Every output falls out of simple arithmetic you can verify by hand. There is no AI, no opaque estimator, no lookup table keyed off a vendor. The goal is a budget you can defend in a finance review, not a quote.
The five inputs, in plain English
Product engineers. The headcount you are building the platform for. Do not include the platform team itself or DevOps consultants. This is the denominator of your cost-per-developer-served metric.
Loaded cost per platform engineer. Blended annual fully-loaded figure. Salary plus benefits, employer payroll taxes, recruiting amortisation (annualised over 3-4 years), laptop and home-office budget, and per-head SaaS subscriptions. The default of $190k reflects a senior-weighted team in US metros. Use $160k for a remote-first or regional mix, $220k for Bay Area heavy.
Target ratio. How many product engineers each platform engineer serves. Industry data clusters between 1:8 and 1:12. Lower (fewer product engineers per platform engineer) is more service-oriented; higher is more self-service-dependent. The default 1:10 is a reasonable middle.
Tooling tier. Spend per product engineer per year across the seven tool categories a platform team typically procures. Lean means open-source-first, free CI tier, minimal observability. Standard means mainstream commercial tooling across the stack. Enterprise adds compliance features, premium support, and scaled compute.
Cloud model. Managed (cloud-native managed services for your Kubernetes, observability backbone, state store) has higher fixed cost and lower per-engineer variable. Self-hosted saves on licences but costs more engineer-time per unit of scale.
Worked example: 150 engineers at defaults
At 150 product engineers with the default settings ($190k loaded cost, 1:10 ratio, Standard tooling, Managed cloud, overhead included) the model returns a platform team of 15 engineers, $2,850,000 in salary, $750,000 in tooling, $102,500 in cloud, and $427,500 in overhead. That is $4.13M per year all-in, or around $27,500 per product engineer served.
That cost-per-dev figure is higher than the $11k-$18k sweet-spot band because the default ratio of 1:10 produces a platform team that is slightly oversized for 150 engineers, and standard-tier tooling plus managed cloud both sit above their low-end bands. To get into the sweet spot you could slide the ratio to 1:12 and switch to Lean tooling, which drops the total to roughly $2.8M and the per-dev cost into the healthy range.
This is why the calculator matters: it makes the trade-offs between team size, tooling tier, and cloud model visible and quantified. Most organisations are not over-spending on platform engineering in the aggregate; they are over-spending on one specific input that the calculator surfaces in seconds.
What it does not model
Three things the calculator deliberately leaves out. First, the value side. ROI requires attribution of engineering-time savings to the platform, which is measured differently in every organisation; /roi covers the framework. Second, the hidden costs like documentation debt, migration double-running, and build-vs-buy churn. These are 15 to 25 percent on top of the visible budget; /hidden-costs has the inventory. Third, non-US salary bands. The default loaded cost reflects US market data. For UK or EU teams, scale the loaded cost down by 25 to 40 percent.