Anthropic Restricts Access to Claude Amid Growing Pressure

Anthropic Faces Growing Demand

Anthropic has recently decided to adjust the usage limits of its AI model, Claude, in response to increased pressure on its computing resources. This decision comes after the Pentagon blacklisted Anthropic, which has sparked renewed interest in the company and its technologies.

The changes affect subscribers of the free, Pro, and Max tiers of Claude. While the weekly caps remain unchanged, users will find their five-hour session limits reached more quickly during peak hours, defined by Anthropic as 5 AM to 11 AM Pacific Time, or 1 PM to 7 PM GMT. This means that users will need to manage their usage more carefully during these times.

Impact on Users and Adjustments

Thariq Shihipar, involved in the development of Claude, clarified that despite efficiency improvements, about 7% of users, particularly those at the Pro level, will hit session limits they would not have reached before. He advised moving token-intensive tasks to off-peak hours to maximize sessions.

Shihipar also mentioned that the company continues to invest in effective scaling and will keep users informed of progress. This strategy aims to better manage the growing demand for Claude while ensuring optimal use of available resources.

Industry Context and Challenges

Dario Amodei, CEO of Anthropic,’s refusal to grant the Pentagon unlimited access to the company's AI models has heightened public interest in their technologies. However, like its competitors, Anthropic faces growth challenges related to increased consumption of computing resources.

The emergence of tools like OpenClaw, an open-source AI agent, allows users to fully leverage the capabilities of AI models, thereby increasing resource consumption. Meanwhile, OpenAI has decided to discontinue Sora, an AI video generation application, to refocus its resources on its core services.

At the New York Times Dealbook Summit, Dario Amodei expressed his reservations about the strategies of some competitors, highlighting the risks associated with reckless management of computing resources. Tech giants like Microsoft and Google continue to invest heavily to maintain their position in the AI race.