Odau Report 01 · 2026

The stack changed shape.

SaaS did not disappear. Cloud and AI layered on top. The bill now arrives through more systems, owners and pricing models.

Six minutes · Every number sourced · Nothing gated

Sixty seconds

305

Average SaaS applications

Almost flat year over year

Zylo, 2026

+8%

Average annual SaaS spend

While app count barely moved

Zylo, 2026

98%

FinOps teams managing AI

Up from 31% in 2024

FinOps, 2026

39%

Report any EBIT impact

Use remains ahead of value

McKinsey, 2025

The surface looks stable

The number of apps barely moved. The cost did.

Zylo’s 2026 benchmark held the average SaaS portfolio at 305 applications while average annual spend rose 8%, to $55.7 million.

The question is no longer only how many tools you have. It is how each one grows.

−0.07%

Change in average application count

+8%

Change in average annual SaaS spend

Source: Zylo, 2026 SaaS Management Index.

The budget is moving

AI costs are surprising more companies.

AI-native application spend rose 108% overall and 393% at organizations with more than 10,000 employees. ChatGPT became the most-expensed application.

Seventy-eight percent of 218 IT leaders reported unexpected AI or consumption charges. Sixty-one percent cut projects after unplanned SaaS increases.

78%reported unexpected AI or consumption charges

61%cut projects after unplanned SaaS increases

Source: Zylo, 2026 SaaS Management Index and accompanying survey of 218 IT leaders.

The visible number is incomplete

The first price is rarely the final cost.

IBM and FinOps research point to the same budget problem: operating costs keep moving after approval.

A purchase price is not a complete forecast.

What leaders know

AI costs continue changing after launch.

What leaders need

A complete number before they commit.

Sources: IBM, Cost of Compute; FinOps Foundation, FinOps for AI.

The work crossed the old boundaries

The teams managing the cost have already crossed the old boundaries.

In 2024, 31% of FinOps respondents managed AI spend. In 2026, 98% did.

AI98%

SaaS now or planned90%

Licensing64%

Private cloud57%

Data center48%

Labor28%

Source: FinOps Foundation, State of FinOps 2026. 1,192 respondents representing more than $83 billion in annual cloud spend.

Use is not value

Adoption is ahead of accountability.

Eighty-eight percent regularly use AI somewhere. About one-third have begun enterprise scaling; 39% report any enterprise-level EBIT impact.

Use is not value. The business result still has to be proven.

88%regularly use AI in at least one business function

~⅓have begun scaling AI across the enterprise

39%report any enterprise-level EBIT impact

These are three findings from the same survey, not a conversion funnel.

Source: McKinsey, The State of AI in 2025.

What we repeatedly saw in the field

What published research cannot tell you

A benchmark cannot explain your company.

Research shows direction. It cannot tell you what your exact exposure is, why it changed or what decision should come next.

Your number has to come from your environment.

Benchmark

Shows the direction of the market.

Your company

Requires an answer grounded in your environment.

A recurring pattern across environments our team worked in—not survey data.

Six questions leadership should be able to answer.

These test whether the business has a decision—not how the answer is produced.

0 of 6

If those answers are difficult to produce, the stack has changed faster than the way it is governed.

Sources and method

This report keeps two forms of evidence separate. Published surveys and benchmarks remain attached to their original sources. Odau field evidence comes from recurring patterns our team encountered through direct operating experience; it is identified separately and is not presented as survey data.