GitHub has announced a pause on new subscriptions for Copilot Personal to address the computational pressure caused by AI agent workflows. Product Vice President Joe Binder noted that AI agents require prolonged, high-concurrency computing, far exceeding the original system design expectations. To ensure service stability for existing users, GitHub has implemented restrictions on new users and introduced session limits and weekly token allowances, gradually transitioning to a token-based consumption model while removing or increasing the cost of using certain high-end models. This phenomenon reflects a widespread resource bottleneck in current AI infrastructure, with similar restrictions also being adopted by Anthropic, Google, OpenAI, and cloud providers AWS and Azure, highlighting the structural mismatch between the large-scale deployment of large models and the underlying computational supply.
Author and source: AIBase
Due to unprecedented pressure on AI computing resources, GitHub, owned by Microsoft, has officially stopped accepting new individual subscriptions for Copilot. The world’s largest code hosting platform is caught in a struggle between budget constraints and service commitments.
Joe Binder, GitHub’s Vice President of Products, noted that AI agent-driven workflows have revolutionized computing demands. These “agents,” capable of autonomously executing complex tasks, typically require prolonged, high-concurrency computing support, consuming far more resources than originally anticipated.
Hashrate shortage leads to service degradation
To maintain service stability for existing users, the platform has had to implement restrictions. Without limiting new user growth, all developers risk a significant decline in service quality.
In fact, this shortage of computing power has become an industry-wide issue, with major players such as Anthropic, Google, and OpenAI previously tightening usage restrictions. Cloud service providers are also struggling, with even giants like AWS and Azure frequently encountering capacity bottlenecks recently.
Shift in billing model and benefit adjustments
To address high operational costs, GitHub has adjusted its usage policies by introducing stricter session limits and weekly token caps. This means developers must wait for the window to reset before continuing usage after reaching their consumption limit.
In addition, the official plan is to gradually phase out the fixed-rate model and transition to a token-consumption-based billing system. As part of the cost-reduction initiative, some expensive high-end models have been removed from the subscription scope or billed at higher multipliers.