Lensa AI went viral because its Magic Avatars gave people a personal result they wanted to post, creators posted first, and the app already had a paywall ready. It did not make $8 million a day. Appfigures estimates a peak of more than $2 million net in one day, and $37.5 million net over 60 days to January 2023.
Video length 10 min 41 sec | Published on YouTube January 30, 2023 | Watch on YouTube
Key takeaways
- Our video title overstated the revenue: Appfigures puts Lensa's peak at more than $2 million net in a day, not $8 million.
- Lensa earned an estimated $37.5 million net in the 60 days to 6 January 2023, about 91% from iPhone owners.
- The spike was sharp: 232K downloads in one day, up to 3x its prior peak, from an 8K daily average the month before.
- It faded fast: a group of AI photo apps including Lensa fell from 4.3 million daily downloads to about 952,000 within two months.
- The repeatable parts are the share mechanic, a paywall in front of the feature that costs money, and a listing and onboarding that were ready before the wave.
- AI apps monetize well: RevenueCat reports revenue per install above $0.63 after 60 days for most AI apps, double the all-app median.
What the video covers
- Why Lensa went viral (0:00)
- From Prisma Labs to Magic Avatars (1:52)
- Viral growth engines (2:49)
- Personalisation (4:53)
- Riding the AI wave (6:15)
- Monetization (6:53)
- Paid ads (7:51)
- Social media and UGC (8:43)
How much did Lensa AI actually make?
Lensa AI did not make $8 million a day. The best public data we could verify comes from Appfigures, which estimates app revenue after Apple and Google take their fees. It put Lensa's peak at more than $2 million in a single day, and at $37.5 million of net revenue over the 60 days to 6 January 2023. Appfigures also reported that Lensa's total net revenue had crossed $8 million in early December 2022 and was up to $25 million about a week later. The $8 million we said in the video does not match any daily figure we can verify. The nearest $8 million in the record is that running total. Lensa earned a great deal, fast, but at about a quarter of the daily rate in our title, and as a spike rather than a steady run rate.
The video went up in January 2023, while the numbers were still moving and the wave was still on. Estimates from tracking firms were scattered around the web, and the headline figure we used does not hold up against the firm's own write-ups. We checked it again for this page and corrected it. Treat any viral-app revenue figure the same way: find the tracking firm's original post before you repeat it.
We could not find a verified current revenue number. We found no revenue disclosed by Lensa's owner, Prisma Labs, and the later annual and monthly figures we found on aggregator sites conflict with each other, so we do not quote them. The table below shows only what we opened and confirmed.
| Claim or metric | What the source says | Source |
|---|---|---|
| Video title: "$8 Million/Day" | Not supported. No daily figure near this. | See rows below |
| Normal day before the spike | Net revenue around $16K | Appfigures, 2 Dec 2022 |
| First spike day | Net revenue nearly $160K | Appfigures, 2 Dec 2022 |
| Total net revenue, early Dec | Crossed $8M, then $25M a week later | Appfigures, 16 Dec 2022 |
| Peak single day | More than $2M net | Appfigures, 6 Jan 2023 |
| Last 60 days to 6 Jan 2023 | $37.5M net, about 91% from iPhone owners | Appfigures, 6 Jan 2023 |
| December 2022 ranking | 7th on the top-earning list, about $26M net | Appfigures, 27 Jan 2023 |
All of these are estimates, and all are net of store fees. Gross consumer spending is higher. When you read any app revenue number, check whether it is gross or net before you compare it to anything.
How did Lensa AI go viral so fast?
Lensa went viral because it combined three things at once: a result people wanted to show off, a low-friction way to get it, and a public figure or two posting first. Users uploaded selfies and received stylised avatars of themselves, which is a personal result that is hard to resist sharing. Appfigures recorded 232K worldwide downloads in a day, up to three times its previous peak, when the trend began, after a prior-month average of about 8K a day, and the app reached #2 on the US App Store chart. The avatars were sold as paid packs for a small fee, so trying it felt cheap. Then creators and celebrities posted their avatars, and each post was an advert that cost Lensa nothing. The mechanism is repeatable: ship an output that is personal, visual and flattering, and make the share one tap.
Notice what is not on that list: a big ad budget. Lensa was not a new app: it was an existing photo editor that became the vehicle for the new feature. That is the quiet advantage. The viral feature rode on an app that already had a listing, a paywall, a payment flow and a store ranking. When the wave arrived, there was nothing to build.
The result also spread beyond Lensa. Appfigures counted that the top five AI art apps other than Lensa reached about 11 million installs from late November, with App Store downloads for that group up 1,265% and Google Play downloads up 759%. A trend that big lifts every app in its category, so being first with a polished version mattered.
What is a viral growth engine, and can a normal app build one?
A viral growth engine is a loop in which using the product gives someone a reason to bring in the next user. It has four parts: an experience people love, a built-in reason to share, a low-effort way to share it, and a landing experience that converts the people who arrive. The loop works only if the people arriving also become users who share. Most apps do not need a Lensa-sized loop. They need a small one that improves on paid growth by a few points. That might be a referral reward, a shareable result, a leaderboard or a recap people post each year. Nielsen's 2021 study of 40,000 consumers found that 88% most trust recommendations from people they know. A share from a friend lands with more trust than any ad can buy, which is why the loop is worth building even at small scale.
In the video we used well-known examples. The table groups them by the mechanism, so you can pick the one that fits your app.
| Mechanism | Example | What a smaller app can copy |
|---|---|---|
| Reward for referral | Dropbox gave extra storage for invites | Give both sides something the app already makes cheap to give |
| Streaks and leaderboards | Duolingo | One streak counter and one weekly ranking; see our Duolingo streak teardown |
| Personalised output | Spotify yearly recap, Lensa avatars | Turn the user's own data into something worth posting |
| Lock-in through bundled value | Amazon Prime | Bundle features so the paid plan feels like one decision |
| Feed personalisation | TikTok, Netflix | An interest-based starting feed, chosen in onboarding |
For the full playbook on referral mechanics, see our app referral program guide. A viral loop also needs a place to land. If a referred friend arrives and hits a slow onboarding, the loop leaks at its widest point. Our app onboarding guide covers that step.
Why did Lensa charge for the AI feature from day one?
Lensa put the paywall in front of the one feature that cost money to run and that people wanted most, instead of giving it away to build a free base first. Generating avatars uses real compute, so every free user is a direct cost. Appfigures noted that Lensa's onboarding ended on a paywall that started a free trial, and that avatars were sold for a small fee on top. The pattern also matches what the wider market shows. RevenueCat's State of Subscription Apps 2025 reported that most AI apps see revenue per install above $0.63 after 60 days, roughly double the $0.31 median for all apps. People pay for AI features that do something clear. The lesson for your app is to price the part that has a unit cost, and test the paywall early rather than after you have a free audience that expects everything for nothing.
Lensa mixed two models: a subscription for the editor and one-off purchases for avatar packs. One-off packs suit a feature people use in bursts. Subscriptions suit a feature people use every week. If your app has both, charging for both is not greedy. It matches price to usage. We cover the trade-offs in our app monetization strategies guide.
There is a retention risk to price in. RevenueCat found that nearly 30% of annual subscriptions are cancelled in the first month. A viral spike brings people who came for one thing. Plan what they see after the first purchase, or most of them leave with the trend.
Did Lensa use paid advertising?
During the viral spike Lensa leaned on organic sharing, but it was not a stranger to paid acquisition. Appfigures observed that before the trend, Lensa appeared to use Apple Search Ads across many keyword combinations around erasing backgrounds, a classic way to buy installs from people already searching for a solution. In the video we also described a test run by Lensa's parent company, in which it compared advertising a paid subscription directly against advertising a free trial, using a conversion lift test (a randomised test and control comparison). The reported finding was that ads for the paid subscription brought in more paying users than ads for the trial. We could not verify the figures behind that test, so we give you the idea, not a number: test the offer in the ad, because the answer can run against your intuition. For the channel basics, see our Apple Search Ads guide.
The sensible order is to use the organic wave to learn which message works, then put paid spend behind the version that converts. The wave gives you free data on the creative and the offer. A paid campaign without that data is a guess.
What happened to Lensa after the wave?
Lensa fell back quickly, as most trend-driven apps do. Appfigures noted in early January that the app was still earning about five times more per day than in November, but that its run looked over. TechCrunch, reporting on a group of AI photo apps that included Lensa, found a mid-December peak of more than 4.3 million daily downloads and about $1.8 million a day in consumer spending. By 12 February 2023 the same group was at around 952,000 downloads and $507,000 a day. By our arithmetic that is a fall of about 78% in downloads and 72% in spending in under two months. A viral wave is a spike you convert into a base, and the conversion has to happen while the spike lasts. The apps that keep a base after a trend are the ones that give the new users a reason to open the app next week.
In the video we noted that Lensa's Instagram account shifted after the hype from feature tutorials to reposting user-made content, with users tagging the app to be featured. That is a sound move. It turns customers into a content supply, and it gives them a small reward, being featured, for posting again.
The wave also brought criticism, about how avatars depicted some users and how uploaded photos could be used. If your viral feature handles faces, voices or personal data, write the privacy terms in plain words before you launch, not after the first headline.
What we see across 300+ apps
We have shipped 300+ apps since 2015, with 52M+ installs between them. A few patterns repeat when an app catches a trend, and none of them are about the trend itself.
- The wave rewards apps that are ready. The apps that gain from a spike already have a store listing that converts, a working paywall, and onboarding that does not leak. If you have to build those during the spike, you miss it. Our App Store Optimization guide covers the listing side.
- Virality is a multiplier, not a strategy. A loop multiplies whatever retention you already have. If day-one retention is weak, a bigger loop just fills a leakier bucket faster. Fix retention first. Our app retention benchmarks and the free retention benchmark checker help you see where you stand.
- Charge for what costs you money. AI features have a per-use cost. We would rather test a paywall in week one with a small audience than learn in week twelve that a free tier is burning the budget.
- Pair the organic spike with paid learning. Use the creative that is working organically as the starting brief for paid. This is the cheapest way to find an ad that works.
- Do not build your plan around luck. Trends are real, but you cannot schedule one. You can schedule the work that lets you use one: listing, paywall, onboarding and share mechanics.
If you are planning how to market an AI app, our AI in app marketing guide goes further.
Frequently asked questions
How much did Lensa AI make per day? Appfigures estimated a peak of more than $2 million net in a single day, after store fees. Its estimate for the 60 days to 6 January 2023 was $37.5 million net. We could not verify any figure near $8 million a day, so the number in our video title was wrong.
What is Lensa's revenue today? We could not verify a current figure. We found no revenue disclosed by Prisma Labs, the company behind Lensa, and later annual and monthly numbers on aggregator sites conflict. The last well-sourced data we found is Appfigures' early 2023 estimates, so we use those and no later figure.
Who owns Lensa AI? Lensa is made by Prisma Labs, the company that earlier built the Prisma photo filter app. Lensa began as a photo editor and became widely known when it added its AI avatar feature, called Magic Avatars, in late 2022. We found no financial results published by Prisma Labs.
Was Lensa's viral growth luck? Not entirely. The product gave users a personal result they wanted to share, early posts came from people with large audiences, and the app already had a paywall and store presence ready. The timing, with public interest in AI art high, was fortunate, but the setup that let Lensa use it was not.
Can a small app copy Lensa's viral marketing? Yes, at a smaller scale. Pick one share mechanism, such as a personalised result, a referral reward or a streak, and make sharing one tap. Fix onboarding and retention first, because a loop multiplies what you have. Expect a modest lift that improves paid results, not a sudden spike.
Did Lensa's popularity last? No. TechCrunch, reporting on a group of AI photo apps including Lensa, showed daily downloads falling from a mid-December 2022 peak of more than 4.3 million to around 952,000 by mid-February 2023. Trend-driven apps need a reason for users to return, or the base fades with the trend.
Related
- Our app marketing service
- App referral program guide
- Mobile app marketing case studies
- Video: the Duolingo streak teardown
- AI in app marketing
Planning a launch or a growth push for your app?
Tell us where the app is and what you are trying to move. We will say what we would do first.
Book a strategy callSources
- Lensa AI Is an Old New Photo Editor That's Topping the Charts Right Now, Appfigures, 2022. https://app.appfigures.com/resources/insights/20221202?f=4
- Lensa AI and the AI art app trend (15 December 2022 insight), Appfigures, 2022. https://appfigures.com/resources/insights/20221216/amp?f=1
- Is the Ride Over for Lensa AI (and AI-Generated Art)?, Appfigures, 2023. https://appfigures.com/resources/insights/20230106/5-is-the-ride-over-for-lensa-ai-(and-ai-generated-art)
- Lensa AI Slipped In: The Highest Earning Apps in December, Appfigures, 2023. https://appfigures.com/resources/insights/20230127/4-lensa-ai-slipped-in-the-highest-earning-apps-in-december
- The AI photo app trend has already fizzled, new data shows, TechCrunch, 2023. https://techcrunch.com/2023/02/13/the-ai-photo-app-trend-has-already-fizzled-new-data-shows
- The State of Subscription Apps 2025, RevenueCat, 2025. https://www.revenuecat.com/blog/company/the-state-of-subscription-apps-2025-launch.md
- Nielsen study reveals top attributes for successful adverts (Trust in Advertising 2021), Research Live, 2021. https://www.research-live.com/article/news/nielsen-study-reveals-top-attributes-for-successful-adverts-/id/5093662