Scaling the 'Cursor for Slides' to $50M ARR ft Gamma founder Jon Noronha

Gamma founder Jon Noronha explains how a near-failed startup reached $50M ARR and 50M users by solving the blank-page problem before ChatGPT existed.

• Claude wins on creative taste; Gemini Flash wins on cost efficiency; reasoning models (including DeepSeek) actively hurt creative output.
• PowerPoint and Google Slides each have 500M monthly active users; Gamma’s 50M users are still a tiny fraction of the addressable market.
• Gamma runs hundreds of AB tests per model release, measuring edits, exports, shares, and free-to-paid conversion — not just ratings.
• Fine-tuning is rejected: it hobbles model intelligence and open-source models don’t match closed-source quality for their use case.
• Canva reached $500M ARR without hiring a single salesperson, a PLG benchmark Gamma explicitly aims to follow.
• Gamma uses ~20 image models concurrently (Ideogram and Flux are primary) and A/B tests them the same way it tests LLMs.
• An API launch is imminent, targeting workflows like auto-generating pitch decks from CRM data without salesperson effort.
• Gamma survived SVB collapse and dwindling runway; image models (Stable Diffusion, DALL-E) — not LLMs — were the original ‘why now’ insight.

2025-08-19 · Watch on YouTube