Founder Case Study: Cursor (Anysphere): December 2021, MIT Dorm Room

The four of us were huddled around a single laptop, completely stuck. For six months, we’d been building an AI tool for mechanical engineers working with CAD software—computer-aided design. The problem? None of us knew anything about mechanical engineering.

Aman was staring at a visualization of a piston assembly. Sualeh was reading documentation about gear ratios. Arvid was running calculations that looked like ancient Greek to me. We were brilliant at building AI systems but utterly ignorant of the domain we were trying to disrupt. The project was failing, and we all knew it.

Then Sualeh said something that would change everything: “Why are we trying to solve problems for an industry we don’t understand? We’re software engineers. We live and breathe code. Let’s build something for ourselves.”

That night, we pivoted. We decided to build an AI tool for the one thing we actually knew deeply: writing software. We had no idea that this dorm-room realization would lead to the fastest-growing startup in history—from zero to $100 million ARR faster than any company ever, then to $1 billion, then to $2 billion, all in under three years .

By November 2025, we had raised $3.4 billion from Accel, Coatue, Thrive Capital, a16z, Google, and NVIDIA . Our four MIT-founder team, all under 30, each held stakes worth over $1.3 billion . And it all started because we admitted we were building the wrong product for people we didn’t understand.