The Graphite Acquisition: Why It Explains Cursor Origin

On December 19, 2025, Cursor announced a signed definitive agreement to acquire Graphite, a code review platform Cursor's own announcement described as used by hundreds of thousands of engineers. Graphite's calling card is stacked pull requests: splitting one large change into small, dependent PRs that review faster, managed through its gt CLI, with a merge queue and an AI reviewer (Diamond) built around the workflow.

Context makes the price signal sharper. TechCrunch framed the deal as the third in an acquisition run, after the CRM startup Koala in July 2025 and recruiting firm Growth by Design that November, with Cursor itself valued at $29 billion as of November 2025. Paying a premium over a $290 million valuation for a review tool is a statement about how central the company thought review and merge would become.

That price signal only carries so far, in two directions. Terms were undisclosed, so the premium is Axios's reporting rather than a figure anyone can check against a filing. And a premium tells you how highly Cursor rated review and merge at the end of 2025, not whether the forge its team went on to build works.

In December the deal read as a tuck-in: an editor company buying a review tool. Cursor's announcement carried the tell, though. It argued that "as writing code has become faster, reviewing changes, merging them safely, and collaborating effectively have increasingly become the bottlenecks to building production-grade software," and teased integrations plus "some more radical ideas we can't share just yet."

The radical idea surfaced on June 16, 2026, when Tomas Reimers walked on stage at Compile and demoed Origin, a git hosting platform built by the Graphite team inside Cursor. Set against that demo, the premium starts to look like the price of the people who would go and build a forge, rather than the price of a review tool. That is an inference, and Cursor has never put it that way. Graphite's own announcement did name the direction: "one integrated platform where humans and agents create, review, and merge code changes collaboratively."

Six months from announcement to stage demo is quick for a git host, and Cursor has not said how much of Origin predated the deal.

The strategy, as far as anyone can read it off two announcements and one demo, is a pipeline that keeps every stage of a change under one roof: write code in Cursor with agents doing more of the typing, review it with the patterns Graphite proved (small stacked diffs, a machine review pass before humans), and host plus merge it on Origin. Each Origin component has a different amount of history behind it, and that difference is what you should plan around.

The takeaway for anyone weighing Origin sits in that last column, and it is more qualified than a column can be. Graphite has run stacked review and a merge queue for years at real companies, so the review model itself is proven. It is proven on top of GitHub, though, where the pull requests and the hosting belong to someone else, and inside Origin those same patterns would sit on a host the team is writing from scratch. Years of production use is strong evidence about the model and much weaker evidence about the new code underneath it. A comparison of GitHub's stacking, Graphite and Origin makes the split visible feature by feature.

Both companies answered directly, and in writing. Cursor's announcement: "Graphite will continue to operate independently with the same team and product." Graphite's own post: "Graphite's product and brand aren't going anywhere," with continued investment promised in the stacked PRs platform and merge queue, plans to combine Graphite's AI reviewer with Cursor's BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition., and integrations connecting local development, background agents and pull requests through 2026.

For teams using Graphite today, nothing needs migrating. The gt workflow still runs against GitHub, reviews and the merge queue work as before, and neither company has announced a feature removal since the deal. The open question is positioning, and neither company has published anything about how Graphite and Origin are meant to sit together once Origin ships.

The live question for anyone already using it is whether to invest further while the same people build a host that would replace GitHub. The instinct is to wait for Origin. Waiting costs more than it saves here: stacks and the queue produce ordinary GitHub pull requests underneath, so the layer comes off later if the positioning changes, and in the meantime you get the review workflow you were going to want anyway. A five-person team on one repository has less to weigh either way, since stacking pays off mainly where review is the bottleneck.

The promise worth checking is the specific one: combining Graphite's AI reviewer with Cursor's BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition.. If that ships, the integration story held up. The cheaper signal is whether Graphite keeps shipping at all, and it would show up first.