Cursor just hit $3B in annualized revenue, and SpaceX holds a $60B option to acquire it.
Cursor, the AI coding tool that did not exist three years ago, just hit three billion dollars in annualized revenue, and SpaceX is reportedly holding a sixty billion dollar option to acquire it. That combination of numbers tells you something important about where the AI industry is putting its money in late 2025: not into the models themselves, but into the tools wrapped around them.
The revenue story alone is worth pausing on. Over three thousand enterprises are now paying Cursor more than one hundred thousand dollars annually. That is not the shape of viral consumer adoption. It is the slow grind of procurement cycles, security reviews, legal redlines, and budget approvals, the full enterprise motion that usually takes years to build. Cursor compressed it into roughly thirty six months, which is almost unheard of for a developer tool.
What is actually driving the spend is worth understanding clearly. Cursor is not autocomplete with a nicer interface. It is an agent layer embedded directly inside the development environment, reading across an entire codebase, maintaining context over thousands of lines, and executing multistep tasks with limited supervision. The underlying language model is mostly a commodity input at this point. The value sits in the scaffolding, the editor integration, the context management system that decides what to feed the model, and the workflow patterns that determine how a developer and an agent actually collaborate on real work.
This is the shift that matters for the broader field. The competitive battleground in AI has moved from raw model benchmarks to the tooling layer that sits on top. Foundation model labs that only sell API access are watching margin migrate downstream, into the IDE, the agent harness, and the enterprise contract. Anthropic and OpenAI both clearly see this, which is why they are pushing harder into their own coding products. The application layer is finally where the durable economics live.
SpaceX's interest fits the same logic from the buyer side. When you employ thousands of engineers and your competitive advantage depends on shipping software faster than anyone else, owning the tool they use every day is not a distraction from the core mission. It is vertical integration into your own productivity stack, the same instinct that led Amazon to build AWS and Apple to design its own silicon.
The question worth watching is what happens to Cursor as a product if the acquisition closes. A tool used by three thousand outside enterprises and one used exclusively by a single rocket company are very different things, and the answer will say a lot about whether the next phase of AI infrastructure stays open or quietly gets absorbed inside the largest engineering organizations.
Originally posted on LinkedIn.