Lensa AI (Prisma Labs)

1. 2022-10

   Baseline before avatars

   219,000 downloads in October

   [2]

2. 2022-11-21

   Magic Avatars launch; number one on the iOS App Store Photo & Video chart

   [2]

3. 2022-12

   Consumer spending from launch to 12 December (about three weeks)

   "consumers have spent $15.4mn worldwide on Lensa AI since the magical avatars launch on November 21, 2022"

   [1]

4. 2023-01

   The fall (secondary data)

   1.4M monthly downloads (December: 19.3M)

   [5]

5. 2023

   Annual revenue (secondary data)

   2022: $39.24M; 2023: $18M

   [5]

Acquisition channels

One channel, zero cost: the user's own face. Avatars became Instagram and Twitter profile photos; every share was an ad. Downloads rose 631 percent month over month in November 2022, with Brazil alone at 31 percent of downloads [2].

Loop

Upload selfies, get a pack within minutes, make it your profile photo, a friend asks, downloads, back to the start. The loop is perfect for a new user but a single lap for the same user: there is no reason to buy avatars a second time (opinion).

Activation

Payment comes before value: the pack is paid for, generation follows. Conversion is high because the social pressure is immediate; everyone's profile shows a new face at the same time (opinion). $15.4M of consumer spending in roughly three weeks, from the 21 November launch to mid-December [1].

Retention evidence

Negative evidence, from secondary data: 19.3M downloads in December 2022, 1.4M in January 2023; annual revenue $39.24M in 2022 and $18M in 2023 [5]. The product rested on one feature; the feature was a moment.

Role of pricing

A one-off pack collected the money of the viral moment immediately. According to an insider quoted by NPR, compute was about half of revenue [3]; so every sale was tied to GPU and gross margin stayed near half regardless of the size of the wave (opinion).

Distribution moves

No distribution move; the wave came on its own. The model (Stable Diffusion) was open source [4]; the feature was easy to copy, there was no moat (opinion).

Team

Around 30 percent of the team was laid off in March 2022; the company had to move its headquarters to Cyprus [3]. The size of the growth team is unknown.

The five questions of the Growth Claim Check, applied to this company's public claims.

Test your own claim with the same five questions: Growth Claim Check →

• Reports of non-consensual and sexualized images generated from selfies of women and girls; a gender bias not present when processing images of men (December 2022) [4].
• Privacy: the terms gave the right to use photos to operate and improve the app; the company said photos were deleted after the model was trained [2].
• The same team's 2016 Prisma wave faded in a similar way; two waves, two fades [3].
• The total revenue figure a company official gave NPR conflicts with Sensor Tower data; it is not used in this archive [3].

This section is opinion; a reading, not a source.

1. 01

   (opinion) A viral wave brings revenue, not a company. Tens of millions of dollars in one month, then under a million a month. Calculating a growth rate during the wave is the textbook example of visible growth.

2. 02

   (opinion) Recognize the single-lap loop. If the user has no reason to come back a second time, it is not a loop, it is a campaign. For a product that explodes on a trend, the first question: what is the reason for a second use?

3. 03

   (opinion) Cash from the wave is for what comes after it. During the wave the growth team's only job is to keep the cohort record and prepare the second feature.

4. 04

   (opinion) An open-source model is not a moat. With a model everyone had, dozens of copies appeared within weeks. Context and habit retain, not the model.

5. 05

   (opinion) The size of the wave does not set the margin. If compute is half of revenue, every sale goes to GPU. For a lira-priced, dollar-GPU product the ratio would be worse.

A question for you: remove this company's logo and drop the loop onto your own product. Which node stays empty?

Numbers are the references in the text. An 'opened' source was read by the researcher; 'secondary' ones are verified against their primary source before publication.

1. [1]
   Monday Mobile Memo

   Sensor Tower · 2022-12-12 · opened

   “consumers have spent $15.4mn worldwide on Lensa AI since the magical avatars launch on November 21, 2022”

2. [2]
   Lensa AI climbs the App Store charts as its 'magic avatars' go viral

   TechCrunch (Amanda Silberling) · 2022-12-01 · opened

   “$3.99 for 50 unique avatars (five variations of 10 different styles)”

3. [3]
   A rocky past haunts the mysterious company behind the Lensa AI photo app

   NPR / WUWF (Bobby Allyn) · 2023-01-20 · opened

   “No deal ever materialized.”

4. [4]
   Prisma Labs (compilation)

   Wikipedia · accessed 2026-10-05 · opened

   “In late November 2022, Lensa's 'magic avatars' feature was launched, which, for a fee, uses artificial intelligence and users' uploaded selfies to create portraits.”

5. [5]
   Lensa AI statistics (Statista/AppMagic relay)

   prioridata · accessed 2026-10-05 · secondary

   Note: secondary relay; download and revenue figures for December 2022 and after

• The December 2022 total and the January 2023 drop are secondary relays only; the Statista or Sensor Tower report should be opened before publication.
• The situation after 2024 (is the app still live, what is Prisma Labs doing) was not researched.