ECONOMY June 19, 2026 · 4 min read

Sensor Tower's State of AI 2026: Time Spent Doubles, Revenue Concentrates

Sensor Tower's State of AI 2026 report reveals global time spent on generative AI apps doubling year-over-year and in-app purchases crossing $4 billion in H1 2026 — but three apps capture 90% of that attention. Here's what the concentration story means for app builders outside the top tier.

By the AppsOps news desk · June 19, 2026 · Original source ↗

Sensor Tower published its State of AI 2026 report on June 16 — right in the middle of WWDC week — and the numbers are striking. Global time spent on generative AI apps is on track to hit 36 billion hours in the first half of 2026, up from 17.2 billion hours in the same period last year. In-app purchase revenue from AI apps is expected to cross $4 billion in H1 2026 alone, a 36% jump from H2 2025. For anyone shipping or monetising mobile apps right now, the report is worth a close read — because behind the headline growth there is a concentration story that changes the calculus for everyone outside the top three.

Three apps, 90% of the attention

ChatGPT, DeepSeek, and Google Gemini collectively account for nearly 90% of all time spent on AI assistant apps in Q1 2026. That level of concentration is normally associated with mobile gaming, where a handful of titles absorb most of the playtime. It has now landed in what was historically the most fragmented part of the App Store.

ChatGPT hit one billion monthly active users in May 2026, becoming the fastest mobile app ever to reach that milestone — three years from launch, beating TikTok, Instagram, and YouTube. But the more interesting data point is that ChatGPT's True Audience market share fell below 50% for the first time in March 2026 as Gemini and Claude gained traction. The AI assistant market is simultaneously large and consolidating: growing fast, but rewarding the top player disproportionately while challengers carve out the margins.

For indie developers, this is not a story about competing with ChatGPT. It is a signal about where not to position your AI feature set — and where the actual runway exists.

The non-assistant opportunity: category AI tools

The 200,000-plus apps that now reference AI in their App Store descriptions are tracking toward approximately 10 billion downloads in H1 2026, a 25% year-over-year increase. These are not AI assistants. They are fitness apps with AI-generated training plans, finance apps with AI-powered cashflow analysis, language-learning tools with AI conversation partners, and productivity utilities with AI summarisation layers built in.

This is where the long tail of the AI economy lives. And the revenue backs it up. Subscription revenue across the broader app economy surged 105% year-over-year in Q1 2026, significantly outpacing advertising (+14% YoY) and standard in-app purchases (+29% YoY). According to industry data, non-gaming apps overtook gaming in total consumer spend for the first time — $85.6 billion versus $81.8 billion — driven heavily by AI-adjacent productivity, health, and utility categories.

The pattern is clear: users pay recurring subscription prices for AI features baked into apps they already have intent to use. They do not need a general-purpose AI assistant — they need an AI that understands their workout history, their spending habits, or their vocabulary practice routine.

Claude's ARPU trajectory is instructive

One data point in the Sensor Tower report stands out even if you are not building an AI product directly. According to the report, Claude's average revenue per user (ARPU) in the US mobile market climbed from under $0.50 in September 2025 to $2.76 by May 2026 — nearly a 5x increase in under nine months, achieved while growing an audience, not just upselling an existing one.

What that trajectory illustrates is that being third or fourth in a category does not preclude strong monetisation, if the audience that finds you has high willingness to pay and perceives genuine capability differentiation. The same logic applies to specialised AI apps: a smaller, engaged user base paying $7.99–$12.99 per month outperforms a large free-tier audience with thin conversion.

Four things to do with this data before Q3

Update your metadata if you have AI features. Apps mentioning AI are capturing a meaningful download premium. If you have shipped an AI-driven feature and it is not prominent in your title, subtitle, or first screenshot, you are leaving organic discovery on the table — especially as the App Store increasingly surfaces AI-labelled apps in editorial and search.
Anchor your subscription pricing above $4.99. The AI category has trained users to accept $9.99–$19.99 per month for capable, specialised tools. Apps without AI features are still competing at $1.99–$2.99; AI-featured apps are landing at $7.99 and converting. If you have not revisited your price point in the last two quarters, now is the time.
Go geo-first on AI feature copy. AI feature descriptions tend to be the newest metadata in any app, and newest metadata is typically localised last — or not at all. If your AI summary, AI coaching, or AI analysis feature copy only exists in English, you are invisible to the majority of your potential market. If you are unsure about the cost of fixing that, run the numbers on our localization cost calculator.
Consider purchasing-power pricing for AI tiers. Time-spent growth on AI apps is accelerating in APAC and LATAM, but those markets are price-sensitive. If you are offering a single global AI subscription price, you may be leaving conversion on the table in high-growth, lower-GDP markets. PPP-adjusted pricing can close that gap without cannibalising premium markets.

The bottom line

The State of AI 2026 report is a useful reality check. The engagement numbers are real, the subscription revenue momentum is real, and the structural shift from gaming to non-gaming is real. But the concentration at the top of the AI assistant market is also real. Three apps hold the assistant space. The rest of the opportunity is in deep vertical tools that serve a specific user job better than a general model can — and that has always been where indie developers and small mobile teams have won on the App Store. AI does not change that dynamic. It accelerates it.