MyFitnessPal, the two-decade-old calorie-tracking app used by tens of millions of people, closed its acquisition of Cal AI in December 2025 — a deal that handed the veteran nutrition platform control of a two-year-old rival built largely by teenagers and already generating more than $30 million in annual revenue. The purchase is as much a statement about where AI is taking consumer nutrition apps as it is a conventional roll-up of a competitor.
A Teenager’s App Outgrows Its Makers
Cal AI was co-founded by Zach Yadegari, who built the app in high school with his friend Henry Langmack and is now 19 and attending college while still running the business day to day. The app’s premise is simple compared with MyFitnessPal’s manual-entry-heavy legacy workflow: users photograph their meal, and Cal AI’s computer-vision model estimates the dish’s ingredients, portion sizes, and calorie and macronutrient content automatically, cutting out the tedious search-and-select process that has long been the biggest usability complaint about calorie tracking. In under two years, the app logged more than 15 million downloads, an adoption curve that caught the attention of an incumbent MyFitnessPal had been quietly tracking among roughly 70 competitors it monitors in the category.
Why MyFitnessPal Bought Rather Than Built
MyFitnessPal CEO Mike Fisher has been candid about what drove the deal, saying of Yadegari: “You have a conversation with them…and you walk away saying this is an impressive young man,” adding, “this is someone who’s not doing this as a hobby. They’re really serious about it.” Financial terms of the acquisition weren’t disclosed, but the strategic logic is straightforward: rather than spend years trying to replicate Cal AI’s photo-recognition accuracy and viral growth engine internally, MyFitnessPal opted to absorb the technology and its seven-person team plus contractors while keeping the Cal AI app running independently alongside its own, now plugging Cal AI’s camera-based logging into MyFitnessPal’s roughly 20-million-item food database for improved accuracy.
A Summer of AI Features Across the Category
The Cal AI deal landed alongside MyFitnessPal’s own Summer 2026 product push, which added an in-app AI Coach offering on-demand nutrition guidance, a new Plan tab, deeper progress insights, and support for tracking GLP-1 medications like Ozempic and Wegovy — a feature set responding directly to how many users’ relationship with food logging has changed now that appetite-suppressing drugs are mainstream. Competitors have moved in parallel: apps like Fitia have rolled out 24/7 AI coaches that build weekly meal plans and critique diet quality, underscoring that photo-based logging and conversational nutrition coaching have become table stakes rather than novelties across the category in 2026.
Two Products, Two Philosophies
Notably, MyFitnessPal has chosen to keep Cal AI running as a separate product rather than folding it entirely into the parent app, acknowledging the two serve different user instincts: Cal AI is built for speed-focused users who want a near-instant photo-to-calorie estimate and tolerate some margin of error, while MyFitnessPal’s core app remains oriented toward accuracy-focused users willing to manually verify entries against its larger, more precisely curated food database. That bifurcation reflects an open question across the AI nutrition app market — whether users ultimately prioritize frictionless speed or verified precision, and whether a single company can serve both without diluting either experience.
The Skeptics’ Concerns
Nutrition and eating-disorder specialists have raised recurring concerns about the broader category of AI-powered calorie-counting apps, warning that frictionless, highly accessible calorie tracking — especially tools fast enough to log every bite via photo — can exacerbate disordered eating patterns in vulnerable users by reinforcing obsessive food monitoring. Computer-vision calorie estimation also carries real accuracy limitations: portion size estimation from a single photo is inherently imprecise, cooking methods and hidden ingredients like oils or sauces are difficult for a model to detect visually, and errors can compound for users relying on the numbers for medical reasons, such as diabetes management. MyFitnessPal has not detailed what, if any, additional safeguards it plans to build into Cal AI post-acquisition.
What’s Next
The deal sets up a test of whether incumbent-scale food databases paired with viral-growth AI tools from younger challengers can coexist profitably under one roof, and whether MyFitnessPal’s bet on camera-first logging pays off against competitors racing to add similar computer-vision features. With GLP-1 tracking, AI coaching, and photo-based logging all converging as standard features in 2026, the next differentiator in the category is likely to be accuracy and trustworthiness of the underlying AI models rather than any single flashy capability — a race MyFitnessPal is now trying to win by owning two competing approaches to the same problem at once.
Photo: Daria-Yakovleva / PIXABAY via Pixabay
