Skip to content

Blog

Cal AI's $50M MyFitnessPal exit: deal terms + what changes

By Alec Zakhary

TL;DR

MyFitnessPal acquired Cal AI in a deal closed December 2025 and announced March 2, 2026. Founders Zach Yadegari (19) and Henry Langmack bootstrapped to $50M ARR in 18 months. Post-deal, differentiation moves to USDA accuracy (Cronometer) and restaurant specificity (Nutrogine).

Illustration for Cal AI's $50M MyFitnessPal exit: deal terms + what changes

The deal that everyone in the calorie-tracking space had been quietly watching closed in December 2025 and was announced on March 2, 2026: MyFitnessPal acquired Cal AI. It is the most consequential consolidation move in this category in years, and the founders are 19 and 20 years old.

I am writing this from a particular angle: I am a product manager building a different calorie-tracking product (Nutrogine, restaurant-first, ships Q3 2026) and I am explicitly outside the existing app industry. I have no equity in either side of this deal. What follows is what I think the move actually means for the market, the users, and the alternatives — based on public reporting, my own research, and a fair amount of reading other people’s reaction takes.

What actually happened

The mechanics are surprisingly clean for a deal this size:

  • Closed: December 2025
  • Announced: March 2, 2026 (TechCrunch broke it; press release went out the same day)
  • Acquisition price: undisclosed
  • What MyFitnessPal got: Cal AI, the photo-scan app, with ~15M cumulative downloads, ~30 employees, and revenue numbers that depend on which source you read (see footnote)
  • What Cal AI got: access to MyFitnessPal’s database — 20 million food entries, 68,500 brands, 380+ restaurant chains, plus the corporate parent (MFP itself is owned by Francisco Partners)
  • Post-deal structure: Cal AI continues to operate as a standalone app. MyFitnessPal also ships Cal AI’s photo-scanning capability inside the MFP app

A note on the revenue numbers. Three sources cite three different figures: TechCrunch reports “$30M+ annual revenue” attributed to MyFitnessPal at the time of announcement; Inc. reports “$40M over the past 12 months” in their Yadegari profile a few weeks later; and Yadegari himself tweeted “$50M ARR” the day of the announcement. The pattern is consistent with a fast-growing consumer app: TechCrunch’s $30M is the conservative MFP-quoted figure, Inc.’s $40M is the Yadegari-quoted LTM (last twelve months), and the $50M ARR is the run-rate annualizing recent months. All three are correct depending on how you measure. Realistic 12-month revenue around the deal: $30–40M; run-rate at deal close: ~$50M.

The Inc. profile of Zach Yadegari adds biographical detail: Yadegari is 19, a freshman at the University of Miami, and built Cal AI starting at age 17 with co-founder Henry Langmack. They bootstrapped — no venture funding. The acquisition price was not disclosed, but the Inc. piece notes it was a “high-eight-figure deal,” which puts it somewhere in the $50–100M range. For context, that is roughly 1–2x ARR, below typical SaaS multiples but in line with consumer-app M&A — especially for a category where lifetime value depends heavily on subscription churn and platform fees.

Why MyFitnessPal wanted to do this

The strategic logic is fairly transparent if you have been watching the category. MyFitnessPal had been losing share in the photo-scanning use case for two years. Cal AI’s growth was the loudest signal that “snap a photo, get calories” was becoming the default UX expectation, and the existing photo features inside MFP were widely considered weaker than the standalone Cal AI experience.

Three reasons MFP wanted the buy specifically:

  1. Defensive consolidation. Cal AI was the most credible threat to MFP’s position as the default consumer calorie app. If you can’t outbuild a competitor, you buy them. Francisco Partners (MFP’s owner) has both the capital and the appetite — they bought MFP from Under Armour in 2020 for $345M and have been running an obvious roll-up strategy in adjacent categories.

  2. Photo-scan as a feature, not a product. The consensus inside the category by mid-2025 was that pure photo-scan apps would have a hard time scaling beyond enthusiast users. Most casual users want a database lookup with photo as one of several inputs, not photo-only. By acquiring Cal AI, MFP gets to ship the photo capability as a feature inside their broader app — which is exactly what they did in the post-acquisition release.

  3. Database leverage flowing both ways. Cal AI’s photo recognition was strong but its underlying calorie database was limited. MFP’s database had the breadth (20M entries, 380+ restaurants) but was widely criticized for accuracy (~23% random-sample error rate per a 2024 review I covered in a separate post). Combining the photo front-end with the larger database back-end is meant to produce a better product than either had alone — though I would argue the database accuracy problem doesn’t actually get solved by this combination. More on that below.

Why Yadegari and Langmack sold

The Inc. profile gives you the surface-level answer: the founders are young, they built fast, and an acquisition at this stage de-risks them while letting them work on next projects. Yadegari has been publicly clear about wanting to build something else — “now I’m aiming even higher” was a direct quote in the announcement.

The deeper reason, which the Inc. piece hints at without saying directly, is that consumer app distribution at scale is brutal once you cross a certain threshold. App Store fees, Apple Search Ads inflation, paid acquisition CAC, churn at the 30-day mark — at $40M ARR you have enough revenue to be visible, not enough to be defensible against larger players who can outspend you on every channel. Selling to a category leader who can absorb you into their distribution machine is often the right move at exactly that stage.

The bootstrapped angle matters here too. Without venture funding, the founders had no obligation to push for IPO-scale outcomes. They could optimize for a clean exit at a reasonable price rather than a moonshot. Most VC-backed equivalents would have been forced to keep raising and keep growing. Cal AI’s structure gave them the option to take a fair offer and walk.

What changed for users

The user-facing changes are smaller than the deal headlines suggest:

For Cal AI users: the app keeps working. Onboarding is unchanged. The main visible change is that the food database lookups now query MFP’s database under the hood, so the search results when you tap “edit ingredients” are deeper. Pricing is unchanged for now — Cal AI Premium remains at its previous tier.

For MyFitnessPal users: photo scanning is now a first-class feature inside the MFP app. If you previously used MFP for database lookups and Cal AI for photos, you can consolidate into one app. MFP Premium pricing was not changed at announcement but most analysts expect a price hike within 6-12 months as MFP integrates the new feature stack.

For users of neither: the choice landscape simplified. The “MFP vs Cal AI” decision is gone. The remaining choices in 2026 are basically:

  • MFP (now with Cal AI photo) — broadest free database, photo scanning, weakest accuracy
  • Cronometer — USDA-verified, highest accuracy, no photo, paid Gold tier
  • MacroFactor — for athletes and macro coaching, paid only
  • Niche photo scanners (SnapCalorie, BitePal, etc.) — most are now subscale and likely consolidation targets themselves
  • Restaurant-specific tracking — currently a gap; what we are building Nutrogine for

The April 2026 Apple App Store crackdown

Six weeks after the deal closed and one month after the public announcement, Apple briefly pulled Cal AI from the App Store on April 21, 2026 over multiple violations of App Store guidelines. TechCrunch’s coverage framed it as Apple signalling that the post–DMA / post–Epic ruling enforcement environment is not a free-for-all.

The specific violations reported across TechCrunch, MacRumors, and 9to5Mac:

  1. Bypassing Apple’s in-app purchase flow. Cal AI implemented an embedded payment screen using Stripe to unlock premium features, removing Apple’s IAP option entirely from the checkout. Apple’s Guideline 3.1.1 requires that IAP be offered alongside any external payment link — Cal AI was not offering both.
  2. Deceptive billing design. The subscription paywall presented weekly pricing more prominently than the actual amount the user would be charged (typically annualized). Customers were arriving at confirmation screens expecting a small weekly charge and being billed a much larger annual figure.
  3. Manipulative re-prompts. After a user declined the first subscription offer, Cal AI presented a secondary purchase prompt designed to pressure the user back into a sale.

Cal AI fixed the violations and was reinstated to the App Store within days. The episode is short-lived but it is worth flagging because it sits awkwardly with the post-acquisition framing.

The product Cal AI bought into MyFitnessPal had a paywall and billing design that Apple — after closer review — judged manipulative. That does not retroactively undo the deal, but it does mean the consumer-trust profile of the merged Cal AI / MFP product is now part of the conversation in a way it was not in March. For users of either app, the practical takeaway is to read your subscription page carefully when you upgrade — particularly the “$X/week, billed annually as $Y” framing, which several apps in the broader category use and which Apple is now visibly policing.

This is also a reminder that growth-at-all-costs consumer-app patterns (aggressive paywalls, weekly-pricing decoy, re-prompts after dismissal) are still profitable but increasingly enforcement-risky. We avoided the entire pattern in the Nutrogine waitlist by design — single email field, no upsell, no secondary prompt — but the trade-off is slower growth than Cal AI’s curve. Both can be right; they are different bets.

What the deal does not solve

This is where I disagree with most of the post-acquisition coverage I have read. The consensus take is “MFP + Cal AI = better calorie tracking.” I think the math doesn’t quite work that way.

The fundamental problem with consumer calorie tracking has never been the photo. It is the database. MyFitnessPal’s database is mostly user-submitted, with a random-sample error rate around 23% based on a 2024 review I cite in detail elsewhere. Cal AI’s photo accuracy, however good, queries against that same database when it goes to look up calories. Garbage in, garbage out.

Add the restaurant problem on top: Cal AI scans a photo of a Chipotle bowl and matches it to a generic “burrito bowl” entry, not to the actual Chipotle nutrition data. Even with MFP’s 380+ chain restaurant coverage, the photo recognition layer doesn’t know to look up “Chipotle bowl” specifically — it sees a bowl with rice and chicken and finds the closest text match. The most relevant context (which chain, which build, which modifiers) is lost in translation between vision and lookup.

This is the problem Nutrogine is trying to solve, with restaurant-aware lookup that pulls brand portion data instead of guessing from the image alone. But it’s a problem that the Cal AI / MFP combination doesn’t address — it just makes the photo step prettier.

The other thing this deal does not solve is portion variance. Wells Fargo weighed 75 Chipotle bowls across 8 NYC locations in 2024 and found 14-27 oz weights — a 33% spread on the median, 87% spread between the heaviest and lightest digital orders. No photo scanner can tell from an image whether your chicken portion was 4 oz or 8 oz. The merged Cal AI + MFP product still gives you a single confident calorie number for that bowl. That number is mechanically misleading regardless of how good the photo recognition is.

What this means for the alternatives

The category just consolidated meaningfully. Five practical scenarios for what to use in 2026, depending on what you actually need:

1. You want the easiest UX and don’t care much about accuracy

Stay with MFP (with Cal AI photo). The combination is now genuinely good for casual tracking, and the database is the broadest free option you’ll find. Acknowledge the ~23% error rate; verify anything that surprises you.

2. You want serious accuracy

Cronometer. USDA-verified, no photo, our detailed comparison piece goes into the trade-offs. If your tracking has clinical or athletic stakes, the precision difference matters.

3. You’re cutting / bulking / recomping seriously

MacroFactor. The adaptive macro algorithm based on weight trend + intake is the right tool for deliberate body-composition work. Paid only ($11.99/mo or $71.88/year) — no free tier. Built by Stronger By Science, used by serious lifters. See the comparison.

4. You eat out a lot

This is genuinely a gap right now. MFP knows about chain restaurants but doesn’t surface portion variance. Cal AI doesn’t recognize specific chain dishes. Cronometer doesn’t have most fast-casual coverage. The honest answer is that no current app does this well — which is the gap I’m trying to close with Nutrogine, shipping Q3 2026. Until then, the workaround is to look up the brand-published nutrition for the closest matching default build and adjust by 15-25% for known portion variance.

5. You want to track your photos privately

SnapCalorie or BitePal, both of which have stronger privacy stances than the merged MFP/Cal AI offering (which now sends photos to a Francisco Partners-owned company). Neither is as accurate as the merged product, but if data ownership matters more than calorie precision, they’re worth knowing about. See Cal AI vs SnapCalorie and Cal AI vs BitePal for the detail.

The PM angle

I keep getting asked, after writing about this acquisition: are you worried? You’re building a calorie tracker, the category leader just got bigger, isn’t that a problem for Nutrogine?

The honest answer is no, and the reason is that the consolidation actually makes the differentiation easier to communicate. Pre-acquisition, “we have better restaurant data than MFP” sounded like one nuance among many in a crowded category. Post-acquisition, the market is more clearly two camps:

  1. The default consumer app — MFP + Cal AI photo, broad database, mid accuracy, photo-first UX
  2. Specialized tools — Cronometer for accuracy, MacroFactor for athletes, Nutrogine for restaurants

Specialization wins at the long tail of any consumer category that has consolidated. People who eat out frequently — and that is a lot of people — have a real, specific problem that the merged MFP/Cal AI doesn’t solve, and now there’s a clearer category position for the tool that does.

The other thing the consolidation does is make data-source positioning legible. MFP is the crowdsourced database. Cronometer is the USDA-verified database. Cal AI is the photo-first scanner. Nutrogine is the source-cited aggregator. These are now distinct identities in the user’s head in a way they weren’t 18 months ago.

I’m also taking the bet that the Cal AI integration accelerates the photo-scan UX expectation industry-wide. By Q3 2026 when Nutrogine ships, photo scanning will be table-stakes — not a differentiator. What will matter is what the app does after the photo: whether it just gives you a confident-sounding number, or whether it gives you the source attribution to verify that number. We’re betting on the latter.

What I’d watch next

Three things to watch in the calorie-tracking category over the next 12 months:

  1. MFP pricing. Premium is currently $19.99/mo. Once the Cal AI integration is fully shipped (probably mid-2026), expect a price hike. The acquired-app integration playbook usually unlocks a price increase within 6-12 months.

  2. The other photo scanners. SnapCalorie, BitePal, Calorie Mama, Lose It (which has a basic photo feature) — at least one of them will be acquired or shut down within 12 months. Photo-scan as a standalone product is hard to defend post-Cal-AI consolidation.

  3. The accuracy reckoning. As more peer-reviewed research comes out on AI photo accuracy (the PMC systematic review on image-based dietary assessment is one of several recent ones), expect the marketing claim “92% accurate” to become harder to sustain in plain text. The more honest 60-80% accuracy framing is going to get harder for app marketing teams to avoid. We have a research piece on this specifically.

  4. Apple’s enforcement posture on health-app paywalls. The April 2026 Cal AI takedown is part of a broader pattern — Apple is back to active enforcement on subscription deception in the consumer-health category. Expect more paywall takedowns in the next 12 months across the category, particularly for “$X/week (billed annually $Y)” designs that mislead about actual charge amount. Apps that survive the next round will be the ones with transparent annual-price-up-front paywalls and clear cancellation flows.

TL;DR for tracking decisions

If you currently use Cal AI: stay. The product is unchanged in substance.

If you currently use MFP: try the new photo feature. If it’s better than your old workflow, great; if not, no harm done.

If you’re picking a new app today: the /compare page has a use-case-based buying guide that walks through the choice. Photo logging → MFP or specialized scanner. USDA accuracy → Cronometer. Macro coaching → MacroFactor. Restaurant tracking → wait for us, or use brand apps in the interim.

If you eat at restaurants frequently: the restaurant calorie counter methodology page explains why this is a structurally hard problem and how Source Badges + portion variance disclosure address it. The dish-level pages on this site are usable today, even pre-app-launch.

Sources