85% of organizations call managing cloud spend a top challenge and estimate 29% of IaaS and PaaS spend is wasted (Flexera). 98% of FinOps teams manage AI spend (State of FinOps), 82% of container users run Kubernetes in production (CNCF), and 57% of operators’ last major outage cost over $100,000 (Uptime Institute).
Most statistics pages in this industry quote each other in a circle until nobody remembers where a number came from. This one doesn't: 55+ figures on cloud spend, FinOps, DevOps performance, Kubernetes, reliability, AI infrastructure, rates and security — each linked to the primary source that produced it, labeled with its report edition. Cite anything here freely with attribution: link this page or the primary source, per our content provenance policy.
Key takeaways
- 85% of organizations call managing cloud spend a top challenge, and reported waste rose in 2026 for the first time in five years (Flexera).
- 98% of FinOps teams now manage AI spend — up from 31% just two years earlier (State of FinOps 2026).
- 82% of container users run Kubernetes in production (CNCF), while most containers use under half their requested memory (Datadog).
- 57% of operators' last major outage cost over $100K, and 1 in 5 crossed $1M (Uptime Institute 2026).
How much is spent on cloud in 2026, and is waste rising?
| Figure | Finding | Source |
|---|---|---|
| 85% | of organizations cite managing cloud spend as a top challenge | Flexera 2026 State of the Cloud |
| 29% | of IaaS and PaaS spend is wasted, by organizations’ own estimate — a slight rise after five years of decline | Flexera 2026 State of the Cloud |
| 76% | of large enterprises spend more than $5 million per month on cloud | Flexera 2026 |
| 73% | of organizations operate hybrid (public + private) cloud environments; 83% run active workloads on AWS and 79% on Azure | Flexera 2026 |
| $723.4 billion | Worldwide public cloud end-user spending was forecast to reach $723.4 billion in 2025, up from $595.7B in 2024 | Gartner, Nov 2024 |
| $80 billion | Sovereign cloud IaaS spending is forecast to total $80 billion in 2026 | Gartner, Feb 2026 |
| $17.5 billion | India's public cloud spending is forecast to grow 28.1% to $17.5 billion in 2026 | Gartner, Jun 2026 |
| $6.37 trillion | Worldwide IT spending is forecast to grow 14.2% in 2026 to $6.37 trillion, led by AI-driven data center and IaaS investment | Gartner, Jul 2026 |
What do FinOps teams manage, and who do they report to?
| Figure | Finding | Source |
|---|---|---|
| 1,192 practitioners | representing $83bn+ in annual cloud spend answered the sixth annual survey | State of FinOps 2026, FinOps Foundation |
| 98% | of FinOps teams now manage AI spend — up from 63% in 2025 and 31% in 2024 | State of FinOps 2026 |
| 90% | of FinOps teams manage SaaS spend or plan to within a year; 57% manage private cloud and 48% data center spend | State of FinOps 2026 |
| 78% | of FinOps teams report to the CTO or CIO; only 8% report to the CFO | State of FinOps 2026 |
| 8–10 FinOps practitioners | Organizations managing $100M+ in cloud spend average 8–10 FinOps practitioners plus 3–10 contractors | State of FinOps 2026 |
| 63% | of organizations have established FinOps teams and 71% run a Cloud Center of Excellence | Flexera 2026 |
What these numbers mean in practice — and how a FinOps function actually gets built — is covered in What is FinOps? and our cloud billing consulting page.
How is AI changing DevOps performance, according to DORA?
| Figure | Finding | Source |
|---|---|---|
| 90% | of respondents report using AI at work; more than 80% say it increased their productivity | DORA State of AI-assisted Software Development, 2025 |
| 30% | of respondents report little or no trust in AI-generated code (down from 39% in the 2024 report) | DORA 2025 |
| nearly 5,000 | The 2025 DORA research draws on nearly 5,000 technology professionals and 100+ hours of qualitative data | dora.dev, 2025 |
| 1.5% decrease | In the 2024 report, a 25% increase in AI adoption was associated with an estimated 1.5% decrease in delivery throughput and 7.2% reduction in delivery stability | DORA Accelerate State of DevOps, 2024 |
| +7.5% | The same 25% AI-adoption increase was associated with +7.5% documentation quality, +3.4% code quality and +3.1% review speed | DORA 2024 |
What the five DORA delivery metrics measure, and DORA’s 2024 benchmark bands, are covered in our DORA metrics guide.
How many organizations run Kubernetes in production?
| Figure | Finding | Source |
|---|---|---|
| 82% | of container users run Kubernetes in production | CNCF Annual Cloud Native Survey, Jan 2026 |
| 80% | In the 2024 survey: 80% of organizations ran Kubernetes in production (up from 66% in 2023), 93% were using, piloting or evaluating it, and cloud-native adoption hit 89% | CNCF Annual Survey 2024 |
| 77% | of organizations have adopted GitOps principles; 60% use CI/CD for most or all applications (adoption up 31% YoY) | CNCF 2024 |
| 50% to 42% | Service mesh adoption declined from 50% to 42% between 2023 and 2024 | CNCF 2024 |
| Almost two-thirds | Almost two-thirds of Kubernetes containers run for under 10 minutes; about one-third finish in under a minute | Datadog State of Containers and Serverless, 2025 |
| less than half | Most container and serverless workloads use less than half of requested memory and under 25% of requested CPU | Datadog 2025 |
| 9% to 15% | ARM-based instance adoption rose from 9% to 15% of organizations in two years | Datadog 2025 |
The requested-vs-used gap is exactly what our Kubernetes cost calculator prices out, and what Kubernetes consulting engagements fix.
How much does a major outage cost?
| Figure | Finding | Source |
|---|---|---|
| 57% | of operators said their most recent major outage cost more than $100,000; 1 in 5 crossed $1 million for the second year running | Uptime Institute Annual Outage Analysis 2026 |
| fifth consecutive year | Per-site outage rates declined for the fifth consecutive year; power remains the leading cause | Uptime Institute 2026 |
| about two-thirds | Third-party IT and data center providers account for about two-thirds of publicly reported outages over nine years | Uptime Institute 2025 |
| Nearly 40% | of organizations suffered a major human-error outage in three years; 85% of those traced to procedures not followed or flawed | Uptime Institute 2025 |
| 44% | of organizations had incidents tied to suppressed or ignored alerts; 78% had at least one incident no alert caught — customers found it first | NeuBird State of Production Reliability, 2026 |
| 40% | of engineering time goes to incident management instead of product; 34% report downtime costs of $100K+ per hour | NeuBird 2026 |
Convert any uptime target into its real downtime budget with the SLA calculator; the alert-suppression trap has its own teardown.
How fast are AI adoption and GPU spend growing?
| Figure | Finding | Source |
|---|---|---|
| #1 | AI cost management is the #1 most-needed skillset among FinOps practitioners | State of FinOps 2026 |
| 81% | of organizations use generative AI (up from 72% in 2025 and 47% in 2024) | Flexera 2026 |
| 47% | of large enterprises are standing up AI governance teams | Flexera 2026 |
| 40% | GPU-running organizations increased average GPU spending 40% year over year, reaching 14% of EC2 compute costs | Datadog State of Cloud Costs, 2024 |
| 83% | of container costs are associated with idle resources (54% cluster idle, 29% workload idle) | Datadog 2024 |
| just over 6% | Organizations running GPUs grew from ~4.5% to just over 6% (Oct 2023–Oct 2025), with GPU instance hours roughly tripling | Datadog 2025 |
| 63% | OpenAI's share of LLM usage fell to 63% (from ~75% a year earlier) | Datadog State of AI Engineering, 2026 |
| 70%+ | of organizations now use three or more models | Datadog State of AI Engineering, 2026 |
| Only 28% | of LLM call spans used cached-read input tokens | Datadog 2026 |
| 69% | of all input tokens went to system prompts | Datadog 2026 |
| 84% | of developers use or plan to use AI tools; 51% of professionals use them daily | Stack Overflow Developer Survey 2025 |
GPU cost mechanics — MIG, Spot, batching, quantization — are covered in our GPU cost guide and DevOps for AI startups.
Salaries and rates
| Figure | Finding | Source |
|---|---|---|
| $87,011 | DevOps engineers report a global median salary of $87,011; the US median is $165,000 | Stack Overflow Developer Survey 2025 |
| $103,112 | Cloud infrastructure engineers report a global median of $103,112 ($189,000 in the US) | Stack Overflow 2025 |
| £525 | UK DevOps contractor median day rate: £525 (P10 £385, P90 £700) across 5,698 contract ads, as of August 2026 | ITJobsWatch, Aug 2026 |
| £70,000 | UK permanent DevOps median salary: £70,000, flat year over year, cited in 5.08% of all UK permanent job ads | ITJobsWatch, Aug 2026 |
Consulting rates are a different market from salaries: the full region × seniority × engagement-model picture is in DevOps consulting rates 2026, priced against your own numbers by the DevOps cost calculator.
DevSecOps and AI code security
| Figure | Finding | Source |
|---|---|---|
| Roughly 44% | of AI code generation tasks introduced a risky security vulnerability in Veracode's tests | Veracode 2026 GenAI Code Security Report |
| 56% | the average security pass rate across models is 56%, barely changed from 55% in the first (2025) report | Veracode 2026 GenAI Code Security Report |
| 30% | AI-generated Java is still last by a wide margin, with a mean security pass rate of only 30% | Veracode 2026 |
| nearly 1 in 3 | the best model tested still fails nearly 1 in 3 security tasks | Veracode 2026 |
| 53% | Bigger and code-specialized models are not meaningfully safer: large models averaged a 53% security pass rate vs 51% for medium and small, and coding-specialized models 51% vs 52% for general-purpose | Veracode 2026 |
| 15% | cross-site scripting tasks averaged only a 15% pass rate | Veracode 2026 |
| 28.6 million new secrets | were detected in public GitHub commits in 2025, up 34% YoY | GitGuardian State of Secrets Sprawl 2026 |
| 152% | leaked secrets have grown 152% since 2021 | GitGuardian State of Secrets Sprawl 2026 |
| 5.6% | of public repositories contain at least one secret; internal repos are 6x worse at 32.2% | GitGuardian 2026 |
| 64% | of secrets detected in 2022 remained active four years later | GitGuardian 2026 |
| 81% YoY | AI-service secrets exposed rose 81% YoY to 1.27 million | GitGuardian 2026 |
Why AI-assisted code fails security review at these rates — and the pipeline controls that catch it — is the subject of our vibe-coding security teardown and What is DevSecOps?
Citation policy and update cadence
Quote any figure with attribution to its primary source (preferred) or to this page. To cite a single figure, link to its row: each row has a stable anchor such as #stat-cloud-01. Report editions matter: Datadog's cloud-cost figures are the 2024 edition, CNCF's 80/89/93% figures are the 2024 survey while the 82% production figure is the January 2026 survey, and DORA's 2024 and 2025 findings are kept separate above because the numbers moved. ITJobsWatch rates are rolling six-month data quoted as of August 2026. We refresh quarterly and replace superseded figures in place; unverifiable numbers — however popular — don't get in. InfraZen’s compilation on this page is licensed under CC BY 4.0: reuse it with credit and a link to this page. Each figure remains the work of the publisher credited beside it.
Related: Cloud billing & FinOps consulting · DevOps consulting rates 2026 · Best cloud cost optimization firms · Cloud waste calculator · Cloud bill audit checklist · DevOps & FinOps glossary