Editorial illustration for GitHub Copilot Projects Adds Task Tracking to Code Assistant
GitHub Copilot Projects Adds Task Tracking
GitHub logged 7.38 billion commits in September, five times the volume from August, and most of that traffic came from AI agents writing and committing code on their own. The platform already added more than 3 million CPU cores and 120 petabytes of storage back in August after a string of reliability problems. That wasn't enough. GitHub is now provisioning over 4 million additional CPU cores and 160 PB of storage, and rebuilding its underlying Git architecture to split compute from storage, a change the company says could unlock up to 35 times more throughput.
The timing lines up with GitHub Universe, the company's annual developer conference, where it also rolled out Copilot Fusion, a routing system meant to pick which AI model handles a given coding task based on difficulty, cost, and speed. Fusion ships with three tiers, Standard, High, and Max, and GitHub claims the top tier beats leading frontier models on task completion while keeping inference costs close to baseline.
Jay Parikh, EVP of CoreAI at Microsoft, laid out the reasoning behind the infrastructure push in a blog post published ahead of the conference.
At the conference, GitHub also announced the general availability of Copilot Fusion, an intelligent routing engine that decides which frontier AI models to dispatch based on task difficulty and to balance cost, latency, and reasoning capability.
Why this matters
GitHub's infrastructure math tells its own story: 7.38 billion commits in September, a fivefold jump in a single month, isn't developers typing faster. It's agents committing code around the clock, which is why GitHub needed 4 million cores and 160 PB of storage just to keep up. Copilot Projects is the logical next step once you accept that premise.
If agents are generating this much commit volume across parallel tracks, someone (or something) has to track goals, terminal sessions, and work-in-progress diffs, because no human team lead can manually supervise that many simultaneous threads. For developers, this is a real signal to watch: Copilot is moving from autocomplete toward orchestration, and that changes what skills matter. Knowing how to review and steer agent output may soon matter more than writing it yourself.
For founders and tooling builders, GitHub's capacity bet suggests agent-driven development isn't a niche workflow anymore, it's infrastructure-scale reality. We'd want to see how Copilot Fusion's model routing holds up under that same load before calling this fully solved, but the direction is clear enough to plan around.
Related Reading
Browse all Microsoft news →
Open Source D&B rebuilds 642 million‑business database after AI agents hit limits
Open Source Box CEO Aaron Levie Says AI Will Drive "Biggest Cybersecurity Upgrade
Open Source Study: 6.7% of Deepfake Requests on Hugging Face Targeted Children
Related Topic Claude gains shared context in Excel, PowerPoint; Microsoft adds Copilot Cowork
AI Tools & Apps
Related Topic Microsoft AI Chief: "If It Isn't Safe We Shouldn't Build It
Research & Benchmarks
From Archives Nvidia releases new open transcription model to recognize industry jargon
From Archives FlashMap Flow Mapping Speeds LLM Decoding by 6.14x, Cuts GPU Hours
From Archives Microsoft Offers Time-Limited, Voluntary Buyouts to 60,000 Employees
Business & Startups
From Archives Anthropic Plans USD 269 Billion in Spending With Microsoft, Google, Amazon
Business & Startups
Common Questions Answered
Why did GitHub need to provision 4 million additional CPU cores and 160 petabytes of storage?
GitHub experienced a massive surge in commit volume, with 7.38 billion commits logged in September alone—five times the volume from August—primarily driven by AI agents writing and committing code autonomously. The platform's previous infrastructure expansion in August proved insufficient to handle this exponential growth, necessitating the additional provisioning to maintain reliability and performance.
What is Copilot Fusion and how does it work?
Copilot Fusion is an intelligent routing engine that automatically selects which frontier AI models to dispatch for specific tasks based on their complexity and requirements. The system balances multiple factors including cost, latency, and reasoning capability to optimize performance and efficiency across different coding tasks.
What problem does GitHub Copilot Projects solve for AI-driven development?
Copilot Projects adds task tracking capabilities to help manage the massive volume of code being generated by AI agents operating in parallel across multiple development tracks. As agents commit code autonomously at scale, task tracking becomes essential for monitoring goals, terminal sessions, and maintaining coordination across these parallel operations.
How significant is the commit volume increase GitHub experienced in September?
GitHub logged 7.38 billion commits in September, representing a fivefold increase compared to August's volume, with the majority of this traffic attributed to AI agents writing and committing code independently. This dramatic surge demonstrates the accelerating role of autonomous AI in software development and the infrastructure demands it creates.
Further Reading
- Copilot Agents - GitHub
- Kick off a task with Copilot agents on GitHub - GitHub Docs
- Agent management - GitHub Docs
- GitHub Copilot on GitHub.com - GitHub Docs
- About GitHub Copilot - GitHub Docs
