What Does Your Tech Stack Actually Cost?

Tech stack cost per engineer varies significantly by company stage. Startups typically spend $150 to $400 per engineer per month on tools and infrastructure. Series A and B companies range from $400 to $2,500 per month. Growth-stage companies spend $2,000 to $4,000, while enterprises can reach $3,000 to $8,000 per engineer monthly. These figures include cloud infrastructure, developer tooling, SaaS licenses, data platform costs, and security tooling, but exclude salaries and benefits.

In Zylo's 2026 SaaS Management Index, the median company spends about $9,455 per employee per year on SaaS, against an average of $4,830 per employee in the 2025 index (a median and an average are different measures, so the two are not a like-for-like trend). AI-native SaaS spend rose 108% year over year, and ChatGPT is now the most-expensed application. Spend still varies widely by industry and company size, and the same report finds organizations leave about 36% of their SaaS licenses unused. These figures cover SaaS subscriptions only and exclude cloud infrastructure, developer tooling, and salaries.

The average company runs approximately 305 SaaS applications (Zylo 2026 SaaS Management Index), essentially unchanged on the prior year: the index found app counts down 0.07% year over year while total SaaS spend kept rising. Most of this spend is decentralized: business units control about 81% of SaaS spend while IT directly manages only around 15%, so a large share of subscriptions are purchased outside central procurement (shadow IT). Large enterprises add roughly 21 new applications every month, which is why portfolios keep growing even as teams try to consolidate.

Zylo's 2026 SaaS Management Index finds organizations leave about 36% of their SaaS licenses unused. The waste comes from three primary sources: unused licenses (employees who have seats but rarely or never log in), duplicate tools (multiple teams paying for competing products that serve the same function), and auto-renewed contracts for tools that are no longer actively used.

There is no single right percentage. Deloitte's 2023 Global Technology Leadership Study found the average technology budget was 5.49% of revenue in 2022, up from 4.25% in 2020 and 3.28% in 2016. Sector spread is wide: Deloitte's earlier 2016-2017 Global CIO Survey ranged from more than 7% in banking and securities to less than 2% in construction and manufacturing. Treat these as context rather than a target, and size your own budget from the stack you actually run.

At seed stage (1 to 5 engineers), a lean stack typically runs $800 to $2,000 per month, roughly $150 to $400 per engineer, by leaning on free tiers (GitHub, Linear, Sentry) and keeping hosting under $500 per month. Series A companies (5 to 20 engineers) usually spend $3,000 to $12,000 per month as they start paying for tools that save engineering time. The biggest affordability wins for small teams are avoiding enterprise-tier pricing before you need it, not signing annual contracts too early, and resisting tool sprawl while the team is still small.

Fragmentation shows up as duplicate tools, unused seats, and shadow IT. The average company now runs about 305 SaaS applications and leaves roughly 36% of its licenses unused (Zylo 2026), and business units control around 81% of SaaS spend, so much of it is purchased outside central procurement. The direct financial cost is the licenses paid for but not used, plus duplicate tools bought by different teams. The hidden cost is the engineering time lost switching between and integrating tools that do not talk to each other.

The most effective cost reduction strategies start with visibility. Conduct a full stack audit to identify unused licenses, duplicate tools, and underutilized services. Right-size cloud resources that are over-provisioned for their measured CPU and memory use. Negotiate vendor contracts during renewal windows with competitive quotes in hand. Consolidate overlapping tools where possible. Commit to reserved instances or savings plans for predictable workloads; the discount depends on term, payment option and instance family, so read it from your provider's pricing page. Finally, implement ongoing governance with monthly cost reviews and approval workflows for new tool purchases.