Anthropic Source Code Leak and Pentagon Standoff

The evolving landscape of Anthropic and its flagship AI tool, Claude Code, during a volatile period in 2026. This era is marked by a massive financial surge, including a $380 billion valuation and rumors of an imminent IPO, even as the company faces performance degradation and user backlash. Technical reports and social media discussions highlight a reliability crisis caused by internal engineering missteps, such as reduced reasoning effort and memory regressions in the Opus 4.7 model. Furthermore, the sources describe a major security leak that exposed over half a million lines of source code, revealing unreleased features and hidden mechanics. In addition to these internal challenges, the texts examine geopolitical tensions arising from a dispute between Anthropic and the Pentagon regarding the ethical use of AI in military operations. Ultimately, these documents paint a picture of a frontier AI lab struggling to balance rapid growth and high-stakes innovation with operational stability and security.