Cal AI: The Calorie-Counting App Two 17-Year-Olds Built to $30M a Year and 15M Downloads, Sold to MyFitnessPal — a Year of Talks, Price Undisclosed

The history of calorie-tracking apps is a history of fighting input friction. Search for the food, weigh the portion, log every meal — this friction is what makes most users drop out within a month. Take a photo of your meal, and AI estimates the calories and macros and logs them, an app that collapsed all that input friction into a single camera shot was built in May 2024 by two high schoolers, both 17 at the time. Less than two years later, Cal AI had reached 15 million downloads and $30M+ in annual revenue, and was acquired by MyFitnessPal, the incumbent of calorie tracking. The acquisition closed in December 2025 and was announced in March 2026. The price is undisclosed.

The founders are Zach Yadegari and Henry Langmack, high school classmates, still students at the time of the sale. The team: 7 employees plus a small number of contractors. In a post-sale interview, MyFitnessPal’s CEO cited the fact that the two never missed their Sunday-night stand-up meetings, even while in school, as material for the acquisition decision. Is this a “teenager with a hobby,” or a counterpart who can run a business with discipline?, what the buyer’s due diligence was examining was that operating discipline, as much as the numbers.

The record in numbers

Not a technology company — a distribution company

Cal AI’s core technology, calorie estimation from image recognition, is a commodity achievable with major AI-model APIs. Competitors, including MyFitnessPal, have the same kind of feature. The gap that produced $30M in two years lies not in the product but in distribution.

Cal AI’s strategy was heavy spending on influencer marketing across TikTok and Instagram: have fitness influencers make usage videos and saturate young users’ feeds. The universal concern of dieting, times a value proposition that lands in an instant, “just take a photo”, is a near-perfect match for short-form video. That AI-wrapper contests are decided by distribution is the same structure as Chatbase and Base44’s $80M exit; Cal AI is the consumer-app extreme of the pattern.

Narrative fuel was also built into the distribution design. Yadegari has publicized being rejected by all 15 colleges he applied to, and the anecdote, “the high schooler who built a $30M-a-year app got rejected by every college”, spread again and again as viral material in its own right. A founder whose face and story keep generating exposure that costs no ad budget: in Gen-Z consumer apps, the founder being an influencer is part of the distribution strategy itself.

MyFitnessPal’s CEO describes the segmentation between the two products as: the users who want speed over precision go to Cal AI. The users who want accurate records stay with MyFitnessPal. The acquisition, in other words, was not about acquiring a feature but about acquiring the young casual segment the company had failed to capture, along with the distribution know-how for reaching it. The $30M in annual revenue functioned as proof that the segment exists.

“Sell from your strongest position”

The negotiation history reported by TechCrunch is also worth attention. Talks ran about a year, and MyFitnessPal’s CEO said the Cal AI team “didn’t need to sell, so they were comfortable with the offer.” Growing, profitable, in no trouble if the deal fell through. That state was the source of negotiating power.

This is a general law of exits. A company running out of money gets its sale price beaten down. A company that does not need to sell commands a higher one. The teenage founders executed a textbook-correct sale: negotiate for a year from the peak of growth. Selling at the peak also parallels Skeb’s ¥1B exit.

The seller’s motives are imaginable too. At the stage where a standalone app’s growth starts depending on ad spend and the AI-feature race turns into a head-on collision with the giants, you hand the business to a buyer who values your distribution, at the moment it looks most valuable. The year-long negotiation was also insurance against missing that moment.

Risks and caveats

The case is dazzling, but several discounts apply. The figures come from TechCrunch and Entrepreneur reporting (secondary sources), not audited accounts. AI calorie estimation, meanwhile, has faced constant criticism over accuracy, and as an app handling health information it carried regulatory and trust risk, entering the umbrella of an established brand was also a transfer of that risk. Influencer marketing, for its part, is growth with ad spend paid up front, and the actual profit margin is not public.

There is also the churn question: consumer diet apps are a flagship genre for high subscription churn. The durability of $30M a year depends on the acquisition tap, ad spend, staying open. Selling a business whose growth is bought with advertising is also a deal in which the buyer takes on the question of what remains when the tap is closed.

And the biggest reservation is the moment in which this curve was drawn. “A high schooler making $30M in two years” happened at a singular point where three waves overlapped: the AI boom, the peak of short-form video, and the diet market. Trace the same steps without the waves, and the same result does not follow.

Not their age at 17, but their distance from the feed

The teenage-student angle draws the eye as a story, but structurally it is irrelevant. What worked was not their age but their knowing, as natives of those feeds, what would pop with a young audience. Choosing a proposition that lands in an instant, “just take a photo,” and putting it into fitness influencers’ videos is a call only someone who watches that feed every day can make. Proximity to your market can beat capital. Conversely, assembling the same distribution for a consumer app aimed at a generation or culture you do not belong to is difficult even with money.

From the buyer’s side, too, the price attached to that distance. The audience MyFitnessPal’s CEO described, people who want speed over precision, were customers the incumbent had not been able to reach on its own, and $30M a year served as proof that an entrance to that audience could be built in two years. That they could run a year of negotiations as the side that “didn’t need to sell” was also because that entrance was in their hands. For the environment around AI businesses, see our column: AI business in 2026. For the full list of exit cases, see here.