How Unacademy grew from a YouTube channel to a $3 billion edtech company

Executive overview

Unacademy started as a YouTube channel in 2010 and scaled into India's largest educational company by focusing on test prep, a market with strong retention and high stakes motivation. Post-COVID, online demand collapsed 50%, forcing a painful pivot to offline centres and mass layoffs. The founders rebuilt by slashing burn from $150M/year to near breakeven, then launched AirLearn — a Duolingo competitor — reaching $2.5M ARR in six months with zero paid marketing.

Retention and business model matter more than market size — build something people return to or the economics will break you.

From YouTube channel to test prep platform

• Unacademy launched in 2010 as a YouTube-for-education platform, betting that the top educators would teach online if given the right tools.
• Early growth came from a creator app that democratised content production (screen record, voiceover, no editing skills needed) — analogous to what TikTok did for entertainment.
• The platform scaled to 100+ YouTube channels and 1,000 educators; some teachers grew larger than the founders themselves.
• Test prep became the core product because it is a high-stakes tournament — learners want coaches who have passed the same exams, not AI.
• IP contracts locking educators to the platform were only introduced in 2024; before that, popular teachers could leave and take their audiences.

Surviving the post-COVID collapse

• Online education revenue dropped to ~50% of COVID peaks; the business was losing $150M/year.
• Both founders experienced panic attacks, insomnia, and started smoking — they describe it as moving from Mount Everest to Mariana Trench.
• Response: mass layoffs, operational efficiency, marketing and sales discipline, and building a culture obsessed with EBITDA.
• Burn is now ~$20M/year (mostly capex for 40 offline centres) with $150M cash in the bank — seven years of runway.
• Offline centres compound without extra cost: the same centre generates more revenue in year 3–5 without changing its cost base.

Launching AirLearn as a Duolingo competitor

• The founders had 100+ day streaks on Duolingo but still couldn't speak the language — the product is over-gamified and skips grammar fundamentals.
• AirLearn was built around everything they disliked: proper grammar instruction, masculine/feminine noun distinctions, pedagogy over points.
• Every third user review compared it favourably to Duolingo.
• Growth strategy: zero paid marketing, UGC creators on TikTok generating 50M views/month, dedicated influencer reviews, and product-led growth via K-factor optimisation.
• $2.5M ARR in six months, ranked #5 educational app in the US.

Where AI fits — and where it doesn't

• AI will replace teachers in categories where celebrity educators don't matter: language learning, school homework help, coding tutorials.
• Test prep is different — it is a discipline sport; learners want coaches who have passed the same exams and can instil the right mindset.
• Current LLM hallucination at long context windows is a real barrier to reliable AI tutoring.
• The vision: AirLearn becomes a universal AI tutor, starting with language and expanding from there.

Advice for founders

• Keep the team lean — most AI companies are single-digit headcount.
• Business model is now more important than market size: if there is no retention, the journey will be brutal regardless of the opportunity.
• High-frequency, high-retention models (B2B SaaS, food delivery, language apps) outlast utilisation businesses.
• Build competitors to proven products once you have execution capability — copying a working model in a large market beats pure zero-to-one when you need a reliable win.
• Creators who want to separate income from views should move toward products, D2C brands, courses, or angel investing.