GitHub Swallowed the Stacked PR Category Whole

TITLE: GitHub Swallowed the Stacked PR Category Whole

BODY:

Graphite raised $81 million, grew revenue 20x in 2024, and sold itself to Cursor in December 2025 for a price reported above its $290 million valuation. Six months later, GitHub shipped the core product for free.

That is not a coincidence. That is the sequence.

GitHub's stacked pull requests, now in public preview as of July 30, break large changes into small, reviewable PRs — an ordered series of focused layers that can be independently reviewed and merged in a single click.

GitHub watched a category form, let third parties prove demand was real, and absorbed it into the platform at zero marginal cost.

What Graphite Built, and What GitHub Took

The stacked-diff community — Graphite, Sapling, git-branchless — spent years building third-party tooling around this exact workflow. It is now native to GitHub itself, explicitly wired up for GitHub Copilot and other coding agents via a dedicated gh-stack skill.

The stack map lives inside GitHub's own PR interface, meaning reviewers need nothing installed — no Graphite account, no browser extension, no CLI. That single sentence ends the Graphite sales pitch.

The numbers Graphite posted were real. By 2025 it had over 100,000 total users and more than 500 customer companies, including Shopify, Snowflake, Figma, and Perplexity. All of that validated the problem. GitHub solved it for free.

Why Now

GitHub's chief operating officer Kyle Daigle said in April that code commits are on pace to reach 14 billion in 2026, up from roughly 1 billion in 2025. Pull requests opened by AI agents surged from roughly 4 million in September 2025 to more than 17 million in March 2026 — a 4x increase in six months.

When your platform faces a 14x spike in diff volume in a year, the review queue becomes an infrastructure problem, not a developer experience one.

As agentic coding tools make it cheap to generate large amounts of code quickly, the bottleneck in software delivery has shifted from writing code to reviewing it. Native, agent-integrated stacking enforces decomposition and enables parallel review as defaults. GitHub did not ship this feature because it wanted to be generous. It shipped it because unreviewed diff volume was threatening the platform itself.

The Lock-In Architecture

Copilot generates the code. gh-stack decomposes it into reviewable chunks. GitHub hosts the review. GitHub controls the merge. If you exit any piece of this chain, you exit all of it.

Any developer tool company selling workflow improvement on top of Git now competes against that bundle, for free. Platform wins because it does not need to charge.

In December 2025, Cursor announced its acquisition of Graphite, a startup specializing in AI-driven code review and stacked pull requests. The deal was reportedly priced well above Graphite's previous $290 million valuation. Selling to Cursor was the only exit that did not require competing against the infrastructure you were built on.

Cursor is now owned by SpaceX. Graphite is inside Cursor. And GitHub shipped native stacked PRs anyway.

Second-order: if GitHub owns review, GitHub owns which PRs merge. That is code governance. Governance is leverage.

What to watch: Whether GitHub adds permissions, SLA controls, or audit logging to stacked PR flows at the enterprise tier — that is where the monetization lives. Watch whether Sapling or other platform-native tools copy the integration. Most concretely: watch the volume. If review queues actually break at 14 billion commits, GitHub just bought the market by making the problem visible and shipping the only free solution on the same day.