- OpenAI explores an internal code platform after GitHub outages disrupt workflows
- GitHub’s Azure migration caused configuration problems and intermittent downtime for engineers
- Teams have been unable to commit code for hours during recent outages
OpenAI is reportedly developing an internal code repository platform after frequent outages on GitHub disrupted coding operations for its engineers.
The effort appears to be a direct response to GitHub’s ongoing infrastructure migration to Microsoft Azure, which has introduced configuration problems and intermittent downtime.
Multiple outages over the past several months have left teams unable to commit code for hours at a time, raising concerns about reliability for AI-driven development workflows.
Article continues below
Rising outages and reliability concerns
A mid-year report from GitProtect indicated a 58% year-over-year increase in incidents during 2025, with 17 classified as major, resulting in over 100 hours of disruption.
The current migration has placed GitHub in a split-traffic state, with portions of its services still operating from its legacy Virginia data center, while other services have moved to Azure.
An early February 2026 four-hour outage was traced to an Azure problem, and a separate week-long outage was linked to a configuration change, which GitHub publicly acknowledged as “not yet meeting our expectations.”
These interruptions reportedly motivated OpenAI to explore a more controlled environment for its internal development and AI integration.
Employees familiar with the project said the platform could eventually be offered commercially, potentially bundled with Codex coding agents, although it may remain exclusive to OpenAI.
Developing internal code repositories is a common practice among large technology firms, with Google operating Piper and Meta using Sapling, but none have been commercialized.
OpenAI’s project, if made available to external customers, would represent a rare case of a company competing directly with a major investor.
Microsoft owns approximately 27% of OpenAI and acquired GitHub in 2018 for $7.5 billion.
OpenAI has already introduced ChatGPT features overlapping with Microsoft Office applications, suggesting broader strategic friction.
The move reflects broader concerns about the resilience of coding platforms under the growing strain of AI-driven workloads.
Massive AI data center buildouts are squeezing energy supplies, and the efficiency of internal collaboration tools is becoming more critical.
The planned platform may use advanced networking technologies such as Ultra Ethernet to improve data movement across AI infrastructure and reduce dependency on third-party platforms.
Engineers have reportedly experimented with laptops for programming tasks within this environment, integrating vibe coding tools to maintain workflow continuity despite external service interruptions.
OpenAI has not publicly confirmed the project, but the initiative reveals the tension between platform reliability and the growing demands of AI-powered development.
The platform remains in early development and could take months before any operational rollout.
Via Tom’s Hardware
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.
https://cdn.mos.cms.futurecdn.net/WgYZVv6ucLUnQDJXEGANqc-2560-80.jpg
Source link
