Cursor is known to be the fastest growing SaaS company in history. Yet, because they don’t control the underlying model, I’ve always felt the writing was on the wall. They are an AI tool, not an AI product, and tool companies risk disruption from model providers who can build competing tools on their own platforms.
But Cursor has consistently shown they can deliver the highest quality product. The paradigms they introduce set the standard for how AI integrates into developer workflows. When Claude Code and other TUIs emerged with smoother agentic editing, I wondered how Cursor would respond.
Sure enough, Cursor 2.0 proves they can stay ahead by introducing some of the best UI paradigms for an agentic IDE.
What keeps Cursor going is product innovation, not model innovation. They aren’t trying to win the model game1. And while many worry the writing is on the wall, they continue to set the pace for how IDEs adopt AI.
Addendum #
The more I read their subsequent post on kernels and how they went about training Composer (their new custom model), I’m starting to change my thinking about Cursor and ever so slightly more bullish on the writing not being on the wall for them after all. Their first mover advantage (with a fantastic product) allowed them to collect data that’s unique to the coding IDE use case. This has now given them the ability to train a highly effective custom model that could potentially give the Sonnet 4.5s and GPT codexes of the world a run for their money.