Starting at the scale of 10 million US dollars, those little-known AI products that make big money without much fanfare.
The other day, a16z once again released its "Top 100 Global Consumer AI Applications" list, which is the 7th edition this time, and also the list with the smallest changes in the past 3 years. Compared with previous editions, there are only 11 new products on the Web and App lists this time.
But from another dimension, the situation is completely different. For the first time, a16z released a global Top 50 toC AI application list based on monthly revenue, while previous editions only had rankings by web traffic and App monthly active users.
- Note 1: The revenue data is sourced from the data platform YipitData, which mainly tracks the operating performance of enterprises from the perspective of consumer data such as bank card transactions and electronic receipts, covering more than 12 million US users. This list is ranked based on the amount users paid for AI products via personal bank cards in August 2026, which mainly reflects the payment behavior of C-end users for AI products, and does not equal the total monthly revenue of each company.
- Note 2: a16z excludes products mainly oriented to NSFW scenarios from the list.
In this list, 29 products do not appear on the Web and App lists ranked by usage volume in this edition, accounting for more than half, and only 7 products including ChatGPT, Claude, Perplexity, Suno, Photoroom, Canva, and Notion are on all three lists.
Further observing the 50 "most profitable" products sorted by a16z, you will find that in each category there are products that we are not very familiar with, or that we originally thought "would not perform well" but have surprisingly high rankings, for example, Reface, the face-swapping product that we thought had long been "completely cold", unexpectedly ranks ahead of Photoroom, which has always been very profitable.
In this article, we will take this revenue perspective to sort out these "unfamiliar" products.
I. Image/Design
This year, the popularity of AI video is much higher than that of image/design, but the C-end payment for image/design is still strong. Many familiar names are on the list, such as Canva and Figma, the leading products in the design field, Midjourney, the "king" in the image generation track, Gamma and Photoroom in the AIPPT and product image tracks, Topaz Labs, the image quality enhancement product that was just acquired by Adobe in June, and so on.
We will not go into details about these products, and we will focus our perspective on the other three products: Reface, Evoto, and beautiful.ai.
Reface: The resilient Ukrainian vendor,
Survive by transformation after the face-swapping trend fades
We paid attention to "Reface" back in 2020. At that time, the App made videos by swapping users' faces onto celebrities' faces, which triggered social media dissemination and ranked high on the lists of many countries. However, as the popularity faded, its downloads and revenue dropped significantly.
Six years later, although the short video face-swapping function in the App is still there, it has been moved to the third tab. Instead, the App has added popular features in the past two years such as hugging and kissing, AI Yearbook, dance videos, and holiday costumes. From "pure face swapping" to "entertainment-oriented image and video generation", the transformed "Reface" is still active in the market.
Screenshot of Reface's App product
Just like "Reface", this Ukrainian developer famous for "face swapping" has also completed its transformation. Opening its official website, you can find that there are still 8 active applications. In addition to entertainment-oriented AI image/video products, it has also developed practical tools such as AI tattoo (Ink AI) and AI try-on (Letsy). Its calorie tracking product "BitePal" even reached a monthly revenue of 3 million US dollars earlier this year. (Regarding the calorie tracking product "BitePal", White Whale once published an article titled "Redesign the tool with 'pet raising', an APP earns 3 million US dollars a month". Interested readers can read it for more information)
Evoto: One-click to complete studio-level photo retouching operations
"Evoto" can edit details such as clothing folds and glasses reflections | Image source: Evoto official website
Compared with Reface which targets C-end users, "Evoto" does not focus on generation, but on "precision retouching". It is on the list because of its To Pro positioning and the pay-per-image business model. According to the official website, "Evoto" is positioned as "one-click to complete studio-level portrait retouching". Its core functions include removing glasses reflections, smoothing clothing folds, evening skin tone, whitening teeth, adjusting posture, etc., which meets the demand for portrait retouching in commercial photography scenarios such as weddings, portrait photography, and formal professional photos.
Evoto subscription price
In terms of payment mode, Evoto adopts a "point system". After finishing the retouching and exporting the image, corresponding points will be deducted. Users can get points by monthly/annual subscription or separate purchase, and the price of processing a single photo is equivalent to 5-10 cents.
Beautiful.ai: No fancy tricks, focus on automatic typesetting
Both are AI PPT tools, Gamma's idea is to let users quickly generate well-designed demo content with various styles in one sentence, with various aspect ratios and support for various attachments. While "Beautiful.ai" focuses on stability, produces standardized PPT, and serves enterprises and users who need to make PPT frequently.
"Smart Slides" interface
The core selling point of "Beautiful.ai" is the template library and the automatic typesetting function "Smart Slides", after users select a template and fill in text, images and data, "Smart Slides" can automatically adjust the font size, spacing and text box position, and create and modify charts with one click based on data. When users modify the layout, the product provides a fixed mode that can be applied with one click, no manual typesetting adjustment is required.
The enterprise version of "Beautiful.ai" also has a brand resource library function, which can lock the brand's logo, color matching and font when making different slides, without manual adjustment every time. "Beautiful.ai" says that more than 100,000 enterprise customers are using this product.
II. AI Video
Compared with the relatively rich product forms of image/design products, although AI video products are profitable and many of them have an ARR exceeding 100 million US dollars, their product forms are relatively similar. 5 out of 10 products follow the "model aggregation" path, plus integrating Agent functions, and other products including Kling AI, digital human products HeyGen and Synthesia are also quite profitable.
But what is worth noting is still "TOPVIEW", as well as "Meshy" whose revenue has grown rapidly recently.
TOPVIEW: 13-year Alibaba veteran, making digital humans hold products
From a functional point of view, "TOPVIEW" is actually nothing special. The combination of "model aggregation + Agent" is oriented to scenarios such as short dramas and marketing, but the team behind it and the development path of the product are quite interesting.
Resume of Jensen Wu, co-founder of TOPVIEW | Image source: LinkedIn
Jensen Wu, co-founder of "TOPVIEW", worked at Alibaba for 13 years and served as the product director of Taobao. Therefore, the starting point of "TOPVIEW" is also the e-commerce scenario, with the core functions of product images and live digital humans. However, compared with other digital humans that mainly read scripts, "TOPVIEW" can make digital humans hold products in their hands or even wear them. With this difference, TOPVIEW's customers include many well-known brands such as L'Oréal, Anta, and Anker, and it also completed a US$8.5 million Series A financing in August 2025.
After users upload product images, they can generate images of digital humans wearing clothes by smearing and other methods, and then generate videos based on the images | Image source: TOPVIEW official website
Later, "TOPVIEW" also expanded the "reproduce hit content" function, and launched the short drama tool Drama Studio. According to the official website, the current positioning of "TOPVIEW" is "AI Video Agent for short films and marketing videos".
Meshy: AI 3D has truly moved from "playable" to "usable"
On September 30, the ARR of "Meshy", an AI 3D model generation platform founded by Chinese founder Ethan Hu, exceeded 100 million US dollars, a 100-fold increase in less than two years. At present, Meshy has more than 15 million registered users and more than 3,000 enterprise and institutional customers. In July this year, it also completed a nearly 400 million US dollars Series B financing with a valuation of 1.5 billion US dollars, and the lead investors include IDG Capital, MatrixPartners China, and Monolith Capital.
AI-generated 3D models are nothing new, but they have never reached a "usable" level in the past. The core problem is that although the generated 3D model looks good as a whole, its internal structure is chaotic. Once you need to modify it or make the model move, it is prone to deformation, and it will also cause the color and texture of the model surface to be misaligned, affecting the subsequent production process. Users still need to do a lot of "repair" work in professional software, which is time-consuming and laborious.
In the past year, Meshy has focused on solving these problems, optimizing parameters such as model structure and materials that affect subsequent production, and optimizing workflows for scenarios such as game development and 3D printing. For example, it is compatible with Unity, Unreal, Blender and other software, and adds functions such as automatic skeleton binding and action presets. By adapting to the two vertical scenarios of game modeling and 3D printing, Meshy has obtained real and substantial returns.
