Inside the $162.8 Million Mobile AI Intimacy Market Revenue Concentration, Store Tolls, and the Economics of Digital Romance

A flagship market intelligence briefing on the loneliness economy | Last updated: September 2026

Quick answer: The $162.8 million figure is credible only when defined narrowly: estimated gross consumer spending across 214 romantic and adult/NSFW AI-companion apps on Apple's App Store and Google Play during H1 2026 (Appfigures). It is neither total company revenue nor the full romantic-AI market — web-only services like Candy AI, SpicyChat AI, and Joi AI sit outside this measured mobile cohort. The same cohort has generated $427.3 million cumulatively since late 2022 across 165.3 million downloads. This is a genuinely fast-scaling mobile subcategory with unusually strong spend intensity, extreme revenue concentration (top 5 apps capture 47.6% of H1 2026 revenue), and no verifiable market-wide CAGR.

Editorial note on framing: this report avoids the claim that this is "digital romance's fastest-growing consumer category" — no authoritative cross-category ranking establishes that superlative. It is accurately described as a fast-scaling mobile subcategory with strong, measurable spend intensity.

Executive Finding

The $162.8 million figure is credible only when defined narrowly: estimated gross consumer spending across 214 romantic and adult/NSFW AI-companion apps on Apple's App Store and Google Play during H1 2026. It is neither total company revenue nor the full romantic-AI market, because web-only services such as Candy AI, SpicyChat AI, and Joi AI sit outside the measured mobile-store cohort.

Appfigures estimates that the same mobile cohort accumulated $427.3 million in consumer spending and 165.3 million downloads from late 2022 through H1 2026. H1 2026 alone generated 38.1% of that cumulative spending, showing rapid recent monetisation, but the available public data do not provide a consistent annual series from which a defensible category CAGR can be calculated.

The investable thesis is not simply "AI girlfriends are growing." It is that romantic context appears to produce materially higher spend intensity than general-purpose companionship, while revenue remains concentrated, platform tolls are significant, compute costs are undisclosed, and safety regulation can alter the product overnight.

2. Measured Market

Appfigures identified 214 NSFW and romantic AI-companion apps in the two major mobile stores. They generated $162.8 million in H1 2026 and $427.3 million cumulatively since late 2022, across 165.3 million cumulative installs.

This implies approximately $2.58 of cumulative gross consumer spending per lifetime download for the measured romantic/adult cohort. That ratio is a category-level arithmetic indicator — not ARPU, LTV, or realised developer revenue — because downloads are not unique users and the spending and install cohorts accumulate over different user lifetimes.

A broader Appfigures companion-app dataset had 337 active, revenue-generating apps by mid-2025, including 128 launched during 2025. That cohort produced $82 million in H1 2025, reached 220 million cumulative downloads and $221 million cumulative spending by July 2025, and showed 88% year-on-year download growth in H1.

These two datasets should not be spliced into a single market-growth series. The 2025 dataset covers broad AI companions; the 2026 dataset isolates romantic and adult apps. Their category boundaries differ, which makes a direct H1 2025-to-H1 2026 growth percentage analytically unsafe.

3. Monetisation Trend

For the broad mobile AI-companion cohort, estimated revenue per download rose from $0.52 in 2024 to $1.18 in 2025 — a 126.9% increase. The metric indicates stronger monetisation per acquired install, but does not reveal conversion, retention, or subscriber lifetime value.

Appfigures also found that PG/general companion apps generated $164.8 million in H1 2026 — almost identical to romantic/adult apps' $162.8 million — despite the PG cohort receiving nearly twice as many downloads. Romantic framing therefore appears to generate roughly twice the spend per install, although exact category download totals were not publicly disclosed.

4. Spending Concentration

H1 2026 revenue was led by Zeta at $33.0 million, followed by Tipsy Chat at $15.2 million, ChatBox at $13.0 million, Crushie AI at $8.8 million, and Emochi at $7.5 million.

Together, those five apps generated $77.5 million, or 47.6% of the measured H1 category. Zeta alone represented 20.3%. This is a revenue-rank curve rather than a time series; every plotted value is an exact Appfigures estimate with no smoothing.

The earlier broad-companion dataset was even more concentrated: the top 10% of apps captured 89% of category revenue, while only about 33 apps exceeded $1 million in lifetime consumer spending. The implication is a power-law market in which app supply expands much faster than economically successful supply.

5. Store Dependence

The $162.8 million measure is app-store consumer spending, not net developer receipts. Apple's standard commission is 30%, while qualifying developers with up to $1 million in prior-year proceeds can pay 15% through its Small Business Program; crossing the threshold returns the developer to the standard rate for the remainder of the year.

Applying those fee rates mechanically to $162.8 million produces a theoretical post-store pool of $113.96 million at 30% or $138.38 million at 15%, before taxes, refunds, model inference, moderation, payment disputes, staff, paid acquisition, and creator/licensing costs. These are sensitivity bounds, not estimates of actual category net revenue, because the platform, region, subscription tenure, and developer eligibility mix are undisclosed.

The mobile dataset also omits web-first operators. That creates two opposite distortions: measured category size is understated because web revenue is absent, while mobile developers' apparent revenue overstates proceeds because consumer spending is reported before store deductions.

6. Economic Model

Romantic AI products typically combine recurring subscriptions with usage-based scarcity. The subscription unlocks continuity and premium models; token or credit packs monetise marginal consumption such as long conversations, image generation, voice, video, and explicit role-play. The strongest design creates emotional switching costs through memory, relationship progression, and personalised media — but those same mechanisms raise dependency, privacy, and safety exposure.

A proper company-level unit-economic model should separate:

• Gross billings: App-store plus web consumer payments before refunds and taxes.

• Net revenue: Gross billings after platform commissions, refunds, and indirect taxes.

• Contribution margin: Net revenue less inference, image/video generation, storage, moderation, and payment costs.

• Subscriber economics: Paid conversion, monthly churn, reactivation, average revenue per payer, and payer LTV.

• Acquisition economics: Blended CAC by organic, affiliate, influencer, and paid-performance channel.

• Risk-adjusted LTV: LTV after app-removal, age-gating, model-policy, chargeback, and regulatory scenarios.

The currently public category data support only gross spending, installs, and rank concentration. Claims about gross margin, CAC payback, or LTV:CAC should be rejected unless supplied by the operator or verified in diligence.

7. Capital Signals

Venture funding demonstrates investor appetite for personalised AI, but disclosed rounds often cover broader character or companion platforms rather than romantic AI alone. Character.AI raised $150 million at a $1 billion valuation in March 2023 while reporting no revenue; the capital was intended for model training, compute, and team expansion. Replika's earlier Series A.2 was $6.5 million, bringing total funding at that point to $10.9 million.

No authoritative public database exposes a complete 2026 funding series specifically for AI girlfriends, AI boyfriends, and romantic-chat companions. Publishing a category funding total or CAGR would therefore create false precision. The more defensible investment conclusion is that revenue evidence is becoming stronger while funding disclosure remains sparse and category definitions remain unstable.

8. Demand Context

WHO estimates that around 16% of people globally — approximately one in six — experience loneliness, with adolescents and younger adults showing the highest prevalence. WHO distinguishes loneliness (a subjective gap between desired and actual relationships) from social isolation (an objective lack of ties).

Across OECD countries, 10% report lacking support, 8% report no close friends, and 6% report loneliness most or all of the time over the prior four weeks.

These figures establish the social-connection problem, not the serviceable market for romantic AI. Loneliness prevalence cannot be multiplied by subscription price to create TAM because many lonely people will not adopt or pay, while many paying users may seek entertainment, fantasy, or sexual content rather than loneliness relief.

9. Outcome Evidence

Peer-reviewed research in the Journal of Consumer Research finds that AI companions can reduce momentary loneliness, with "feeling heard" acting as a central mechanism. The studies did not demonstrate persistent long-term psychological improvement beyond immediate interactions.

This distinction matters commercially. A product may optimise session frequency and spend while producing only transient emotional relief. Investors should therefore measure whether engagement supplements human connection or substitutes for it, rather than treating higher time spent as an unqualified positive.

10. Risk Map

• Platform-policy risk: Adult or suggestive content is vulnerable to rejection, removal, age-gating, and payment restrictions.

• Revenue concentration: A small number of apps capture a disproportionate share, increasing winner-take-most dynamics.

• App-store dependency: Distribution and billing tolls can absorb 15–30% before other costs.

• Web measurement gap: Large web-first platforms prevent mobile intelligence from representing the full market.

• Model-cost volatility: Voice, image, and video increase engagement but can damage contribution margin without hard usage limits.

• Dependency and safeguarding: Romantic bonding, always-on availability, and memory can intensify attachment, particularly among minors or vulnerable users.

• Regulatory discontinuity: China's 2026 anthropomorphic-AI rules targeted sustained emotional interaction and intimate virtual relationships for minors, forcing major platforms to remove agent features.

• Privacy exposure: Intimate prompts, sexual preferences, voice, and generated media constitute exceptionally sensitive behavioural data.

11. Investor Diligence Checklist

Investors should require a monthly cohort table covering:

• D1, D7, D30, D90, and month-12 retention by acquisition source.

• Free-to-paid conversion and trial-to-paid conversion by platform.

• Gross and net ARPPU, payer concentration, and refund/chargeback rates.

• Monthly subscriber churn, reactivation, and subscription-tenure distribution.

• Text, voice, image, and video inference cost per active and paying user.

• App-store versus web billing mix and effective platform take rate.

• Organic versus paid installs, CAC, and payback by geography.

• Minor-detection, age-assurance, crisis-escalation, and moderation incidence.

• Memory deletion, export, consent, and intimate-data retention controls.

• Revenue sensitivity to removal of explicit content or tightening of store policy.

The decisive metric is not downloads. It is risk-adjusted contribution LTV divided by CAC, supported by durable retention that does not depend on unsafe emotional escalation.

12. Market Outlook

A defensible 2026 category forecast cannot be produced from the public data. Doubling H1 spending yields a $325.6 million annualised mobile run-rate, but seasonality, new launches, policy changes, and cohort decay make that a scenario — not a forecast.

Methodology and Key Caveats

Figures in this report are drawn from Appfigures data (via IThome, WEEX, and multiple secondary sources), Apple's official App Store Small Business Program documentation, Reuters and BusinessWire (Character.AI funding coverage), Zillionize (Replika funding), WHO's Commission on Social Connection, OECD's Social Connections and Loneliness report, Harvard Business School (AI Companions Reduce Loneliness research), and Decrypt (China regulatory coverage), current as of September 2026. The $162.8 million figure measures H1 2026 gross mobile-store spending across 214 romantic and adult AI-companion apps and explicitly excludes web-only platforms; it should never be presented as annual revenue, company proceeds, total market size, or ARR. No authoritative source provides a complete category funding series or verified CAGR for this segment.

Sources

Market Data: Appfigures (via IThome; WEEX; ElectroIQ; NewsGhana; AIBase)

Platform Economics: Apple Newsroom (App Store Small Business Program); Apple Developer News

Investment Data: Reuters; BusinessWire (Character.AI); Zillionize (Replika)

Public Health & Demographics: UN News (WHO loneliness statistics); WHO (Reducing Social Isolation and Loneliness Among Older People); OECD, Social Connections and Loneliness in OECD Countries

Outcome Research: Harvard Business School (AI Companions Reduce Loneliness)

Regulatory Context: Decrypt (ByteDance/Alibaba China regulatory coverage); Decrypt (Joi AI coverage)

Consumer Coverage: Sedaily (romantic AI companion adoption trends)

Inquiries

General Inquiries
contact@lonelinesseconomy.com

Research & Data Inquiries

research@lonelinesseconomy.com
Partnerships & Commercial Inquiries
Partnerships@lonelinesseconomy.com