I built a $50M AI app in high school (and just sold it for...)

What It Covers

19-year-old Zach Yadegari built Cal AI, a calorie-tracking app using food photos, reaching $30M in 2024 revenue and $5.7M in January 2025 alone before selling to MyFitnessPal. The episode covers his acquisition process, growth tactics, decision-making frameworks, and how he navigated college while running a 30-person company.

Key Questions Answered

• •Acquisition strategy: Build a warm intro network before approaching buyers. Zach targeted 10 strategic companies and had mutual contacts frame the outreach as their own idea, not a cold solicitation. First-round offers came in at roughly 1x annual profit — far below expectations. Accepting that the company might never sell, then refocusing on long-term building, paradoxically attracted the eventual buyer, MyFitnessPal.
• •Growth sequencing: Influencer marketing drove Cal AI from zero to $2M monthly revenue, then plateaued. The next growth layer came from paid performance ads on Instagram, TikTok, and Facebook, scaling to over $1M monthly ad spend in January 2025. The Mr. Beast sponsorship cost $500K, returned roughly $350–400K directly, but generated downstream brand credibility that exceeded the initial loss.
• •Idea validation framework: Before building, identify an adjacent competitor's revenue as a floor estimate for market size. Cal AI was validated by seeing a competing app generating $200M annually. Avoid dismissing ideas because competitors exist — a growing competitor signals an underserved market, not a closed one. Filter ideas by personal experience with the problem and clear viral marketing potential.
• •Expected value decision-making: Apply EV calculations to major business decisions, especially exit timing. Multiply the probability of achieving a future outcome by its value, then compare against a certain present offer. Factor in ongoing risk of the business declining to zero. This removes emotion from decisions driven by fear or greed and produces a more objective sell-versus-hold analysis.
• •Skill stacking over specialization: Deep expertise in one area is less valuable than competency across three complementary skills. Zach combined sufficient coding ability, consumer marketing instincts, and entrepreneurial execution. Six months of focused learning in any skill produces enough competency to hire and manage specialists. Understanding code well enough to lead engineers — without being a top-tier engineer — proved more useful than technical mastery alone.
• •Language and self-belief as execution tools: Zach consciously shifted internal language from conditional to definitive framing — replacing "if I become successful" with "when I become successful." He curated his TikTok feed exclusively to motivational content, which directly sparked the Grindclock app idea after noticing comment patterns requesting motivational alarm features. Treating belief as a trainable muscle, not a fixed trait, correlated with sustained execution through repeated rejection and setbacks.