2026 App economy Consumer wallet

The App Wallet Migrated.
People Are Still Paying — Just Fewer Apps, Bigger AI Subs.

A 2026-only data brief on the consumer app spending shift. Total spend is up, per-user app count is flat, AI Assistants are the fastest-growing category, and non-AI subscriptions are quietly being squeezed into ad tiers and bundles to make room. Every figure below is sourced and dated.

20+ primary sources All 2026 data ~80 KB self-contained Last updated Aug 2026
Summary Headline numbers The wallet migration Category winners & losers Enterprise mirror image Consumer behavior Streaming: the canary AI stack anatomy Six structural shifts Sources

TL;DR — five things the 2026 data actually shows


Headline numbers — what changed in 2026

Six metrics that capture the shape of the shift.

Global app IAP revenue

$167B

+10.6% YoY in 2025. Sensor Tower State of Mobile 2026 (Jan 21, 2026).

Non-game IAP growth

+21%

YoY 2025. Non-game revenue passed games for the first time. Sensor Tower.

AI app IAP revenue

$5B

In 2025, +254% YoY. AI Assistants = fastest-growing category. Sensor Tower.

AI app time spent

36B hrs

H1 2026 projected, vs 17.2B in H1 2025. More than doubled YoY.

Apple IAP order volume

−40%

By early 2026 vs 2021–23 baseline. AOV up 59% to $16.86. Measurable AI.

Active AI subscriptions canceled

47%

Of US consumers canceled ≥1 sub in 2026, up from 31% in 2024. Zuora SEI.

Read this together, not separately. Spend is up, order volume is down, app count is flat. That's not a contradiction — that's the migration. Fewer transactions, bigger dollars, AI-shaped.

The wallet migration — the same dollar, different app

In 2015, for every $1 spent on apps, $6 went to games. In 2025, that ratio flipped to $1 : $0.96 — and the spend that left games went almost entirely to AI, social, and creator tools.

App-store consumer spend: apps vs. games

Spend ratio in app stores (iOS + Google Play, IAP) 2015 $1 apps : $6 games $1 $6 games 2025 $1 apps : $0.96 games — crossover $1 games $0.96 apps Source: Sensor Tower / Attn Economy, Jan 2026.

Where the money went in 2025 (top growth categories)

AI +$3.1B Social +$1.9B Drama +$1.6B Cloud +$1.0B Video +$0.8B

Top-5 fastest-growing IAP categories 2025 (Sensor Tower / Attn Economy).

  • AI Assistants: +$3.1B (+254% YoY). ChatGPT, Google Gemini.
  • Social: +$1.9B (+17%). TikTok, YouTube.
  • Short Drama: +$1.6B (+115%). ReelShort, DramaBox.
  • Cloud Storage: +$1.0B (+31%). Google One, Dropbox.
  • Video Editing: +$0.8B (+71%). CapCut, Splice.

"These aren't entertainment purchases. They're utility, productivity, and identity purchases. People are paying for tools that help them create, store, communicate, and think." — Attn Economy

AI Assistants category growth — multi-axis

GenAI app time spent

+110%

17.2B → 36B hrs YoY (H1 2025 → H1 2026). Sensor Tower State of AI 2026.

GenAI downloads

3.8B

In 2025, +100% YoY. Sensor Tower.

GenAI sessions

1T+

In 2025; AI category became 10th-largest by time spent.


Category winners & losers

Five categories accelerated, five contracted. The split follows utility: anything that does one job gets folded into an AI assistant; anything that's a destination or experience holds.

▲ Winning categories

CategoryMetricChange
AI Assistants+254%IAP 2025 YoY (Sensor Tower)
Short Drama+115%IAP 2025 YoY (Sensor Tower)
Video Editing+71%IAP 2025 YoY (Sensor Tower)
Cloud Storage+31%IAP 2025 YoY (Sensor Tower)
Social+17%IAP 2025 YoY (Sensor Tower)
AI Assistants+426%Time-spent YoY 2025 (Sensor Tower)

▼ Under pressure

CategoryMetricChange
Apple IAP order volume−40%By early 2026 vs 2021–23 (Measurable AI)
Standalone news apps↓↓Daily-active declining 2 yrs (TechDrifting)
Standalone translation↓↓Retention suffering (TechDrifting)
Single-purpose utility appsFolded into ChatGPT/Claude/Notion AI
To-do / habit / flashcardsReplaced by OS-level AI + chat assistants
SVOD price hikes → churn41%Canceled ≥1 SVOD in last 6 mo (Deloitte 2026)

The enterprise mirror image

Zylo's 2026 SaaS Management Index (40M licenses, $75B under management) gives the cleanest read on the same migration inside companies.

SaaS spend per org

$55.7M

+8% YoY. App count flat at 305 (−0.07%). Zylo 2026 SMI.

AI-native spend

+108%

+393% in enterprises >10K employees. ChatGPT = #1 expensed app.

Broader AI category

+181%

Fastest in Zylo's dataset. App Dev +81% (#2).

Unexpected AI/usage charges

78%

Of IT leaders hit unplanned costs in last 12 mo. 61% cut projects.

What the data is telling IT: "AI budget is way up — +108%. But app count is flat. The hot AI-native vendors are winning, and as a cohort, everyone else is losing." — SaaStr summary of Zylo 2026 SMI (Mar 2026).

Top 10 most expensed SaaS apps — Zylo 2026

Of the top 50 most expensed apps in enterprises, 8 are AI-native (Zylo). Expense-based SaaS spend +267% YoY.

1. ChatGPT
#1
2–4. Other AI-native
8/50
5. OpenAI API
#5
Lines of business owning spend
81%
IT directly managing spend
15%

Consumer behavior in 2026

Beyond the spend data — how people are actually using apps.

Adults who use AI chatbots

49%

US adults, up from 33% in 2024. Only 24% daily. Pew Research, Feb 17–23 2026 (n=5,119).

AI users paying for ≥1 AI sub

46%

57% among daily users. Goodwater 2026 Consumer Survey.

Avg AI subs paid per user

4

At ~$66/mo. 24% spend >$100/mo. 14% pay for 8+. Bango, Oct 2025.

Cancel-and-restart AI tools

53%

Of US AI users, churn-and-replace is the default management strategy. Bango.

Replaced ≥1 app with AI assistant

32%

In the past year. 36% expect to in next year. TELUS Digital, Sep 2025.

Substitute, not supplement

43%

Of dedicated-AI-platform users fully replaced prior methods. PYMNTS, 2026.

What the AI subscription dollar actually buys

Average consumer AI subscription stack — 2026 Casual user (Bango) $66/mo ~4 AI tools · Bango Q4 2025 Power user (a16z) $151/mo ~5 tools · a16z Top 100, Mar 2026 SMB stack (Zylo) $100–500/mo Per employee, with overlap · Zylo 2026 Sources: Bango Q4 2025 · a16z Mar 2026 · Zylo 2026 SMI

Concentration inside AI

MetricValueSource
ChatGPT weekly actives900Ma16z, Mar 2026
ChatGPT monthly actives1BSensor Tower, May 2026 (fastest ever)
ChatGPT / #2 (Gemini, web)2.7×a16z, Mar 2026
ChatGPT / #2 (Gemini, mobile)2.5×a16z, Mar 2026
Top 3 AI apps = category time~90%Sensor Tower Q1 2026
ChatGPT mobile consumer spend$2.48B+408% YoY 2025 (CREI)
ChatGPT monthly mobile revenue$338MDec 2025 (CREI)
Claude paid subs YoY+200%a16z, Jan 2026
Gemini paid subs YoY+258%a16z, Jan 2026
Claude US mobile ARPU$2.76May 2026, up from $0.50 Sep 2025

Streaming — the canary

Streaming is the test case for what happens to non-AI subscriptions under price pressure and AI substitution. Flat spend, high churn, fast migration to ad tiers.

US streaming spend / household

$69/mo

Flat YoY. Deloitte 2026 DMT (n=3,575, Oct–Nov 2025).

Canceled an SVOD in last 6 mo

41%

Up from 39% in 2024. Annual churn ~40%.

On ad-supported tiers

68%

Up from 46% in 2024. Will cancel favorite service on +$5 (61%).

Will only sub via bundles

31%

44% Millennials · 48% Gen Z. Bango, Mar 2026.

Ad-tier SVOD share trajectory

US SVOD subscribers on ad-supported tiers (%) 2024 46% 2025 54% 2026 68% Source: Deloitte 2026 Digital Media Trends (Mar 2026 release).
Streaming's "AI substitution" is not the headline story yet — it's price fatigue + ad-tier migration. 73% of US consumers are frustrated by price hikes; 61% would cancel favorite service on a $5/mo increase. AI-driven recommendation + summary features are coming, but the wallet pressure is real and immediate. Deloitte 2026 DMT.

The AI stack — anatomy of a 2026 subscription bundle

How a "power user" actually composes their AI subscriptions — and where the dollars go.

$20 / moChatGPT Plus
$20 / moClaude Pro
$20 / moGemini Advanced
$20 / moPerplexity Pro
$71 / moCursor / Copilot coding

≈ $151/mo total, ≈ $1,812/yr before switching-cost loss (a16z, Jun 2026)

What users actually do with multiple AI tools

Multi-app usage is the norm

~20% of weekly ChatGPT web users also use Gemini in a given week (a16z, Mar 2026).

ChatGPT + Gemini split by task (PYMNTS 2026):

  • Shopping: 60% ChatGPT vs 57% Gemini
  • Entertainment: 65% vs 59%
  • Travel planning: 66% vs 57%
  • Learning / self-improvement: 70% vs 60%

The retention problem inside AI

AI apps carry +41% Year-1 LTV premium vs non-AI ($30.16 vs $21.36 median), but retain 36% worse over 12 months.

RevenueCat 2026 State of Subscription Apps (Mar 10, 2026; 75K apps).

"Consumers are not picking a winner. They're building a toolkit and switching by task." — PYMNTS, Feb 2026


Six structural shifts defining the 2026 app wallet

Shift 1

From many cheap apps to few expensive ones

Same $$, fewer titles. Zylo: app count flat, spend +8%. Apple: IAP order volume −40%, AOV +59%. Measurable AI.

Shift 2

From utility to AI

News, to-do, translation, grammar, flashcards, recipes — folded into ChatGPT, Claude, Perplexity, Notion AI. 32% of US AI users replaced ≥1 app (TELUS).

Shift 3

From single-vendor to multi-tool stacks

Avg 4 paid AI subs at $66/mo. 53% cancel-and-restart. Power-user stack $151/mo. Churn is the default.

Shift 4

From premium to ad-supported in legacy media

Ad-tier SVOD 46% → 68% in 2 yrs. 73% frustrated by price hikes. Consumers protect AI spend, squeeze streaming.

Shift 5

From solo subscriptions to bundles

31% will only sub via bundles (44% Millennials, 48% Gen Z). Disney folding Hulu into Disney+. One bill, multiple products.

Shift 6

From human-shopping to agent-shopping

71% expect GenAI to influence ≥50% of spending decisions (Accenture, n=25,590). 32% ready to hand over purchase decisions. Agentic commerce is the next wallet frontier.

Bottom line: people aren't spending less on apps. They're spending differently — concentrating dollars on AI assistants, churn-restarting AI subs as needed, trading utility apps + ad-supported tiers to afford the new stack, and quietly preparing for AI agents to make purchases on their behalf. The wallet didn't shrink. It got rerouted.

Sources

Every figure above is dated and linked. All 2026 unless noted.

Consumer app spend & category data

Subscription & consumer behavior

Enterprise data

AI market share

Earlier (pre-2026, for context)