Two High Schoolers Built an AI 'Photo Calorie' App — $30M ARR in Two Years, Then Sold to MyFitnessPal

Process

Zach Yadegari was 19 when he sold his second company and was driving a Lamborghini with a "CAL AI" license plate. The AI calorie app Cal AI he built with a high-school classmate hit $30M+ ARR and 15 million downloads in under two years, and in March 2026 it was acquired by industry giant MyFitnessPal. But this wasn't overnight luck — he'd been drilling the same skill since he was 7.

Stage 1 — The start (ages 7→16): self-taught coder with a money-making project by 13

Yadegari's "founder gene" switched on at 7, when his mom sent him to a coding summer camp. From then he went into "devour YouTube" self-teaching mode — not just watching tutorials but DMing programmers and creators online to learn their tricks.

At 13, while peers were gaming, he made a counterintuitive move: go make money off the gamers. He built the gaming site Totally Science, whose core feature helped students bypass their school Wi-Fi blocks to play online games — this "beat-the-system" business earned him $60K a year. At 16, he sold it to gaming company Freeze Nova for nearly $100K.

Stage 2 — Enter Cal AI: a 'snap-and-done' minimal product (2023–2024)

After selling his first project, Yadegari's motive was decidedly un-serious — to impress girls, he hit the gym to build muscle. He quickly realized diet was the key, and that the calorie-tracking apps on the market were "outdated and complex," still stuck in the manual-entry era.

His solution was radically simple: replace manual entry with AI image recognition — snap a photo and know the calories. In summer 2023 he teamed up with Henry Langmack, a friend from coding camp; two high schoolers building an app from scratch, actively seeking advice from successful app founders (including Blake Anderson, who'd built hit apps no-code with ChatGPT). In May 2024, Cal AI officially launched.

Stage 3 — The money logic in three layers: demand positioning × low-price subscription × viral marketing

Cal AI's "how it makes money" breaks into three layers:

• Demand positioning: riding the Western healthy-eating wave, the TikTok "calorie counting" topic had amassed 147 million posts, yet incumbent MyFitnessPal was still on manual entry. The product does one thing — recognize food calories from a photo at ~90% accuracy — minimal, but straight at the pain point.
• Low-price subscription: the app is free to download, with a subscription of just $2.49/mo or $29.99/yr. The ultra-low price makes converting from free "painless." Cal AI earned near-five-star ratings from over 5 million users.
• Viral growth: the founding team spent just $2,000 to test social media, picking health and fitness influencers to naturally embed the app in lifestyle videos. The result: $28K in month one, over $115K in month two.

Stage 4 — $30M ARR and 15M users in two years, sold to MyFitnessPal (2024→2026)

Growth snowballed. By 2025, Cal AI's monthly gross profit was about $1.4M, the team grew from two to dozens, and in under two years the app hit 15 million downloads and $30M+ ARR (Yadegari himself says it broke $50M at one point).

In March 2026, nutrition-app giant MyFitnessPal acquired Cal AI. The logic was clear: they wanted Cal AI's legacy to continue, while Yadegari said "my life's mission isn't to build a calorie-tracking app." Soon after, he'd already started his next venture, Flow — a brand fighting "doomscrolling," whose first product is an alarm you can only silence by physically getting up and docking your phone. From gaming to calories to attention, his pattern is constant: find a group's universal pain point, solve it with the simplest tech, validate/scale/monetize fast, then hunt the next one.

"In the end, age really doesn't matter. Either you're good at what you do or you're not — and the market decides the outcome." — Zach Yadegari (paraphrased from public interviews)

Source: TechCrunch (acquisition) · CNBC · TechCrunch (2025)

Thinking

Insight 1: Turn a "hassle" into "snap-and-done" — the best products often do just one thing

Cal AI does only one thing — recognize calories from a photo — yet it pierced the manual-entry pain MyFitnessPal hadn't solved in over a decade. Having fewer features made it faster, simpler, and more shareable. Find a "hassle" everyone tolerates, use new tech to make it simple, and that's your opening. The more universal and long-suffered the chore, the more it deserves a redo.

Insight 2: A tech inflection is the window to redo an old category

MyFitnessPal was a decade-long incumbent, but precisely because it was big and burdened, it stayed on manual entry. When the "tech inflection" of mature AI image recognition arrived, a small team could bite market share from the giant with a minimal product. Every major inflection (AI, mobile, blockchain…) opens a window to redo an old category — the key is stepping in just as the window opens, before the incumbent reacts.

Insight 3: Low price is a strategy, not a concession

$2.49/mo makes the paywall painless, and with a huge base of free users, Cal AI wins on conversion rate and scale, not on high ticket. When your product acquires customers virally, "low price + high conversion + big volume" usually beats "high price + low volume." Get as many people painlessly using it as possible, then earn on conversion.

Insight 4: Growth via "product-as-content" + influencer seeding, not ad spend

They spent just $2,000 to test, letting fitness influencers naturally surface the product inside real lifestyle videos — $28K in month one. The precondition is that the product is inherently "results-visible and filmable" — photograph a plate, instantly get calories, perfect for a short-video demo. When choosing a product, ask: is it filmable? Worth an influencer casually showing it off?

Insight 5: The serial founder's meta-skill — keep asking "why would people pay for this?"

From gaming site to calories to attention, Yadegari kept answering the same question. What transfers is never a single product but the loop of "find a universal pain point → validate minimally → monetize fast." And this is the part to see honestly: he was self-taught at 7 and running this loop by 13 — what an ordinary person should really learn isn't "build a calorie app," it's turning this method into muscle memory.

Action

Step 1: Find a "hassle everyone tolerates," and use new tech to make it simple

List the tedious things you or people around you endure daily (manual logging, repetitive forms, tool-switching…), and ask each: "Could AI or some new tool make this one tap?" The more universal the pain and the clunkier the existing solution, the bigger the opening.

Step 2: Make the product do one thing, and do it extremely well

Don't pile on features up front. Pick the single most painful action (e.g., "photo → calories"), take it to a 90, and cut the rest. Minimal = instantly understood + easily shared. A product you can explain in one sentence and want to try at a glance beats ten half-built features.

Step 3: Use low price + free download to flatten the paywall

If you can acquire customers at scale, trade a "painless price" (a few dollars a month) for high conversion and big volume rather than high ticket, low volume. Get a huge number of people using it first, then earn slowly on conversion. Free download + an ultra-low subscription is the smoothest ramp from stranger to paying user.

Step 4: Pick a "filmable" product and cold-start with influencer seeding

With a small budget (hundreds to a few thousand dollars), find mid-tier vertical influencers and let them surface your product inside real lifestyle content. The precondition is that the product carries a "makes-you-want-to-try-it" visual — results-visible, a few seconds to demo. Letting content spread itself is far cheaper than hard ads.

Step 5: Drill "find pain → validate minimally → monetize fast" into a repeatable loop

Don't bet on one product. Every time you build, force yourself to answer "why would people pay for this?" — run the loop once, then transfer the method to the next arena. A transferable capability is worth more than any single hit.

Not for you if: you want a "stable" business (viral AI apps are high-variance, mostly-fail territory); or you won't do growth and marketing and just want to code; or you can't make a product minimal enough to explain in one sentence and want to try at a glance.