SaaS Waste Statistics in US 2026 | Unused Licenses, Subscription Costs & Software Sprawl

SaaS Waste Statistics in US

The average organization spends $55.7 million a year on SaaS and wastes about $19.8 million of it on unused licenses, according to Zylo’s 2026 SaaS Management Index. License utilization improved to 54% in 2025 from 47% in 2024, which still leaves about 46% of paid licenses idle.

SaaS Waste in the US 2026 – Introduction

SaaS waste statistics in the US all point to the same problem: companies buy more subscriptions than they use, and they lose track of what they bought. The best 2026 benchmark, Zylo’s SaaS Management Index, covers about $75 billion of analyzed spend. It finds an average portfolio of 305 applications, 33.5% of them bought by employees outside IT, and an average of $19.8 million lost on unused licenses. The federal government shows the same pattern: it spends more than $100 billion a year on IT, and auditors say agencies still cannot reliably compare licenses bought with licenses used.

This report uses vendor benchmarks, not government audits, for most figures, so each table names its source. Zylo sells SaaS management software, and its customers skew toward large enterprises. No benchmark found for this report breaks out US-only results. Summaries of the same data also disagree, with some quoting 36% of licenses unused and others 46%. This report uses Zylo’s own utilization figure and explains the gap in the licenses section.

Interesting Facts about SaaS Waste Statistics in the US

SAAS SPEND VS WASTE PER ORGANIZATION, 2026 ($ MILLIONS)
AVERAGE ANNUAL SAAS SPEND     ████████████████████████████████████████  55.7
MEDIAN ANNUAL SAAS SPEND      ███████████████                           20.6
LICENSE WASTE, 2026           ██████████████                            19.8
LICENSE WASTE, 2025           ███████████████                           20.9
Fact 2026 Data
Average annual SaaS spend per organization $55.7 million (+8% year over year)
Median annual SaaS spend $20.6 million
Average SaaS applications per company 305
License utilization (2025) 54%, up from 47%
Annual cost of unused licenses $19.8 million, down 5.3% from $20.9 million
Shadow IT share of apps and spend 33.5% of apps, 3.7% of spend
Growth in AI-native app spending 108% year over year
Federal IT and cyber spending More than $100 billion a year
Global software spend forecast, 2026 (Gartner) $1.43 trillion (+15.2%)

Waste equals about 36% of average SaaS spend, a share we calculated by dividing $19.8 million by $55.7 million. That share may be the source of the 36% figure some summaries quote as “licenses unused.” Zylo’s own utilization number is 54%, which means about 46% of paid licenses sit idle. The dollar share is lower than the license share, which implies idle licenses cost less than active ones on average, an inference of ours.

The trend is better than the headline. Utilization rose 7 points in a year and waste fell 5.3% even as total spend rose 8%. Costs are shifting, though. AI-native applications cost more, and consumption-based pricing makes bills harder to forecast, so the next benchmark may show waste growing again in a new form.

SaaS Spending per Organization in 2026

SAAS SPEND BY ORGANIZATION TYPE, 2026 ($ MILLIONS)
LARGE ENTERPRISE (AVERAGE)    ████████████████████████████████████████  245
ALL ORGANIZATIONS (AVERAGE)   ██████████                               55.7
ALL ORGANIZATIONS (MEDIAN)    ████                                     20.6
AI-NATIVE APPS (AVERAGE)      ▏                                        1.2
Metric Figure
Average annual SaaS spend $55.7 million (+8%)
Median annual SaaS spend $20.6 million
Average spend at large enterprises About $245 million
Year-over-year spend growth at large enterprises About 16%
Average spend on AI-native applications $1.2 million
AI-native spend growth, all organizations 108%
AI-native spend growth, large enterprises 393%
Gartner global software spend forecast, 2026 $1.43 trillion (+15.2%)

Data Source: Zylo 2026 SaaS Management Index; Gartner forecast as cited by Zylo

The average organization spends $55.7 million a year on SaaS, but the median is only $20.6 million, so a small number of very large buyers lifts the average. Large enterprises spend about $245 million, and their spending grew about 16% in a year, twice the 8% rate for all organizations. Zylo ties much of that growth to AI features and consumption-based pricing.

AI-native applications, meaning apps built around AI, are still a small share of the bill at $1.2 million per organization. They are growing fastest, though: spending rose 108% overall and 393% at large enterprises. Gartner’s forecast of $1.43 trillion in global software spending for 2026 gives the wider context. Only a part of that is SaaS, so use it as a market backdrop, not as a SaaS total.

Unused SaaS Licenses in the US in 2026

SAAS LICENSE USAGE
ZYLO UTILIZATION 2024             ██████████████  47%
ZYLO UTILIZATION 2025             ██████████████████████  54%
VERTICE: FULLY UNUSED (Q2 2026)   ████  14%
VERTICE: UNDER 50% USED           ██████████████████  51%
Metric Figure
License utilization, 2024 47%
License utilization, 2025 54% (a 13% relative improvement)
Licenses unused (our subtraction) About 46%
Average annual waste per organization, 2026 index $19.8 million
Average annual waste, prior index $20.9 million (-5.3%)
Waste as a share of average SaaS spend (our calculation) About 35.5%
Licenses fully unused (Vertice, Q2 2026) 14%
Licenses used at under half of purchased capacity (Vertice) 51%

Data Source: Zylo 2026 SaaS Management Index; Vertice Q2 2026 data as reported by The Finance Circuit

Zylo reports that 54% of licenses were active in the period measured, up from 47% a year earlier. That is a 7-point gain, or a 13% relative improvement, and it cut average waste from $20.9 million to $19.8 million. Vertice, a SaaS buying platform, uses a different test. It finds 14% of licenses fully unused and another 51% used below half of purchased capacity, which adds up to 65% that are unused or underused.

The two sources measure different things. Zylo measures usage against activity windows and recommended utilization levels. Vertice counts how much of the purchased capacity people actually use. Treat the combined picture as a review pool, not as a savings forecast: not every idle license can be cancelled, and renewal terms often limit how quickly a company can cut. Zylo reports that active renewal management programs recover an average of 17% in savings at each renewal cycle.

SaaS Sprawl and Shadow IT in the US in 2026

SAAS PORTFOLIO AND SHADOW IT, 2026
SHARE OF APPS BOUGHT OUTSIDE IT     ███████████████  33.5%
SHARE OF SPEND ON THOSE APPS        ███   3.7%
IT LEADERS FINDING AI APPS/FEATURES
WITHOUT IT'S KNOWLEDGE              █████████████████████████████████  77%
Metric Figure
Average apps per company (Zylo) 305 (median 240)
New apps added per month About 9
Annual app churn rate 33%
Largest enterprise portfolios Up to 1,000 apps
Shadow IT share of apps 33.5%
Shadow IT share of spend 3.7%
IT leaders who found AI apps or features without IT’s knowledge 77%
Average apps per company (BetterCloud survey) 106
Apps not sanctioned by IT (BetterCloud) 44%

Data Source: Zylo 2026 SaaS Management Index; BetterCloud 2026 survey as cited by SaaS research summaries

The average company manages 305 SaaS applications in Zylo’s data, and the median is 240. It adds about 9 apps a month, a 34% annual growth rate by Zylo’s count, while about 33% of apps churn out each year. Employees buy 33.5% of applications but only 3.7% of spend, so shadow IT is mostly low-cost tools. The risk is not the bill. The risk is data, security, and renewal visibility that IT cannot see.

BetterCloud’s survey of IT professionals, which skews toward smaller organizations, finds only 106 apps per company, and 44% of apps not sanctioned by IT. The gap with Zylo comes from method: Zylo counts apps from spend data, and BetterCloud counts what IT teams report. Zylo also finds that 77% of IT leaders have discovered AI features or applications operating without IT’s awareness, which is the newest layer of shadow IT.

AI Pricing and Surprise SaaS Charges in the US in 2026

AI AND CONSUMPTION PRICING IMPACT, 2026
AI-NATIVE SPEND GROWTH (ALL)          ██████  108%
AI-NATIVE SPEND GROWTH (LARGE ENTS.)  █████████████████  393%
IT LEADERS HIT BY UNEXPECTED CHARGES  ███  78%
Metric Figure
AI-native application spend growth 108% year over year
AI-native spend growth at large enterprises 393%
IT leaders who met unexpected AI or consumption charges 78%
Business units’ share of SaaS spend (as reported) 81%
IT’s share of SaaS spend, managed directly (as reported) 15%
Growth in spend through expense channels (as reported) 267%
Median monthly AI token spend (Ramp, April 2026) $2,246
Average monthly AI token spend (Ramp, April 2026) $140,842

Data Source: Zylo 2026 SaaS Management Index; Ramp token spend data as reported by The Finance Circuit

Pricing is changing as fast as usage. About 78% of IT leaders in Zylo’s survey ran into unexpected charges from AI or consumption-based pricing, so a flat per-seat bill no longer describes the cost. Ramp’s April 2026 data shows the skew in AI token spend: the median company spent $2,246 a month while the average was $140,842, so a few heavy users pull the average far above the typical buyer.

Control of the budget has moved too. A summary of Zylo’s data reports that business units control 81% of SaaS spend and IT directly manages 15%, while spending through expense channels rose 267%. Those figures come from a secondary summary, so confirm them against Zylo’s report before you quote them. If they hold, waste is hard to fix centrally, since the people who buy the licenses are not the people tracking their use.

Federal Software License Waste in the US in 2026

FEDERAL SOFTWARE LICENSE MANAGEMENT
ANNUAL FEDERAL IT AND CYBER SPEND ($B)   █████████████████████████████████  100+
SAVINGS FROM BETTER LICENSE MANAGEMENT   ██                                ~2.1
GAO RECOMMENDATIONS TO AGENCIES          ████████████                        18
Metric Figure
Annual federal IT and cybersecurity spending More than $100 billion
GAO recommendations on software license tracking 18 to 9 agencies
Reported savings from improved license management About $2.1 billion
Agency status, early 2026 Some agencies still had not shown consistent tracking
Gartner estimate of unused or underused software 30% to 50% (as cited by a SAM vendor)
Share of federal IT spent on operations and maintenance About 80%

Data Source: US Government Accountability Office, Federal Software Licenses (GAO-24-105717) and Information Technology (GAO-25-107795); Gartner estimate as cited by OpenIT

The federal government spends more than $100 billion a year on IT and cyber-related investments, about 80% of it on keeping existing systems running. In a 2024 review, GAO found that nine large agencies did not consistently track licenses in use or compare them with licenses purchased, and it made 18 recommendations. As of January and February 2026, GAO’s follow-up shows some of those agencies still had not demonstrated consistent tracking.

Better management pays. GAO-linked figures put savings from improved license management at about $2.1 billion. A Gartner estimate that 30% to 50% of software is unused or underused, cited by a software asset management vendor, implies much larger potential waste if applied to federal IT spending, but that is an extrapolation and not a federal audit finding. Congress has also been moving the SAMOSA Act, which would require agencies to review software and plan for unneeded licenses, though it still awaited Senate action when last reported.

Frequently Asked Questions

How much SaaS spend is wasted in 2026?

Zylo’s 2026 SaaS Management Index puts average license waste at $19.8 million per organization, about 35.5% of the $55.7 million average SaaS spend. Waste fell 5.3% from $20.9 million the year before.

What percentage of SaaS licenses go unused?

Zylo reports that 54% of licenses were used in 2025, up from 47% in 2024, so about 46% sat idle. Vertice’s Q2 2026 data finds 14% fully unused and 51% used below half of purchased capacity.

Why do some sources say 36% of licenses are unused?

Several summaries quote 36%. That matches the dollar waste share of about 36%, not Zylo’s license utilization of 54%. Use Zylo’s own utilization number when you describe license counts.

How many SaaS apps does the average company use?

Zylo’s 2026 index reports 305 applications on average and 240 at the median, with the largest enterprises reaching about 1,000. BetterCloud’s survey-based figure is 106.

What is shadow IT in SaaS?

Shadow IT means apps bought outside IT’s control. In Zylo’s data it is 33.5% of applications but only 3.7% of spend, and 77% of IT leaders found AI apps or features operating without their knowledge.

How fast is AI-native software spending growing?

Zylo reports 108% year-over-year growth in AI-native application spending and 393% at large enterprises, with the average organization spending $1.2 million on AI-native apps.

Does the US government overspend on software licenses?

GAO found that agencies do not consistently track software licenses in use against licenses purchased, which leaves cost reductions on the table. Better management has produced about $2.1 billion in savings, and GAO reports follow-up gaps at some agencies as of early 2026.

Are these figures specific to the US?

No. The benchmarks come from global vendor datasets, and this report found no US-only breakdown. Only the federal figures are specific to the US.

How can companies reduce SaaS waste?

Zylo reports that active renewal management programs recover about 17% on average at each renewal cycle. Common steps include reclaiming idle seats, consolidating overlapping tools, and bringing employee-bought apps into procurement.

Disclaimer: This research report is compiled from publicly available sources. While reasonable efforts have been made to ensure accuracy, no representation or warranty, express or implied, is given as to the completeness or reliability of the information. We accept no liability for any errors, omissions, losses, or damages of any kind arising from the use of this report.