Today for AI
HOT RADAR
toolsHEAT 61.4°

AI CLUSTERED EVENT · 10/14/2026

Anthropic Replaces Git with Custom Scion VCS to Handle 30k Daily AI Agent Merges in Five Months

7 reports archived6 independent sourcesupdated 10/15/2026, 04:30:54
Synthesis & Latest Updates
6 Sources Cross-Validated

Anthropic's engineering team completed a five-month migration of its entire Monorepo from Git to a custom version control system named Scion, addressing infrastructure bottlenecks caused by AI agents. As models like Claude autonomously iterate code, daily merge volumes surged 12-fold from ~2,500 to over 30,000, overwhelming traditional Git workflows designed for human-scale development.

Cross-verified reports indicate Scion discards Git's offline-centric 'full copy' mechanism, optimizing instead for high-frequency cloud interactions. It introduces 'Antique Change Requests' to replace Pull Requests, enhancing efficiency for massive parallel submissions. This shift marks a fundamental transition in toolchain priorities from human experience to machine throughput and automated integration.

This event highlights the paradigm shift required in the Agentic Coding era. With coding no longer the bottleneck, CI/CD and version control become limiting factors. Anthropic's case demonstrates that legacy DevOps tools cannot sustain AI-driven extreme throughput, signaling an urgent industry need for native infrastructure adapted to machine-led development scales.

LATEST/Anthropic has solidified Scion as its core infrastructure, sparking industry debate on GitHub's stability under high-concurrency AI loads. As AI-generated code share rises, more tech firms are evaluating custom VCS solutions to handle similar merge pressures, accelerating the shift toward AI-native development toolchains.

HEAT TRENDHourly heat curve

37 fully observed hours
Now
61.4
Peak
96.013:00
24h change
-35%

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

TIMELINECoverage timeline

Total 7 reports · Latest first
  1. Hacker News AIT2·48 pts
    • Scion utilizes a 'derived data' architecture to generate non-changeset data like Blame and Skip Lists via a unified mechanism.
    • The mechanism is recursive, computing child node derived data from parents, allowing easy addition of new acceleration structures.
    • Inspired by Meta's Mononoke, the design aims to optimize specific operations like line-level attribution in large-scale codebases.
  2. InfoQ 中文T2·88 pts
    • Claude's autonomous coding caused daily merge volume at Anthropic to surge from 2,500 to over 30,000, breaking existing Git mirror infrastructures under PB-scale data transfer demands.
    • The new system Scion achieved ~100GB Monorepo migration via fine-grained object storage and incremental sync, improving CI checkout speeds by two orders of magnitude.
    • Engineering paradigm shift: Version control systems are now designed as high-throughput AI collaboration infrastructure rather than just human-interaction tools.
  4. Simon Willison 博客T2·78 pts
    • Anthropic's code merge frequency increased 12-fold in six months, driven primarily by AI agents rather than human engineers.
    • Traditional Git workflows could not sustain 30,000 daily changes, prompting Anthropic to develop a specialized alternative.
    • The new system introduces 'Antique Change Requests' to optimize review and merge efficiency for massive parallel code submissions.
  6. Anthropic AI (@AnthropicAI)T1·78 pts
    • Daily code merges at Anthropic surged from ~2,500 to over 30,000 in months, driven primarily by AI agents rather than human developers.
    • Developer infrastructure designed for human scale became a bottleneck, prompting the company to abandon Git and build a custom version control system.
    • The complete migration of their monorepo from Git to the new VCS was accomplished within five months.
  7. Hacker News AIT2·78 pts
    • Coding is no longer the bottleneck; daily merged changes grew >12x to 30,000+ in six months due to Claude's autonomy.
    • Traditional Git/CI architectures struggle with AI-generated volume, forcing Anthropic to build the in-house Scion VCS.
    • AI-native dev infra focus shifts from optimizing human experience to maximizing machine throughput and automated integration.