Mysterious squad led by a person of Chinese descent safeguards Anthropic's "Glass Wing"

Claim and immediate context

It has been reported that a small, mysterious security squad—allegedly led by a person of Chinese descent—is guarding Anthropic’s Project Glasswing, the public vehicle through which the company this week unveiled the Claude Mythos Preview. The report, published by ifeng, frames the squad as protecting Anthropic’s nascent “Glass Wing” deployment as big‑model competitors ratchet up pressure. These reports are unverified and should be treated cautiously.

Where this fits in the AI arms race

The disclosure comes as the industry roils around next‑generation models: rumors of OpenAI’s GPT‑6, Deepseek‑V4’s cost and context advances, Google’s Gemini pushes, and xAI’s Grok5 leaks. Anthropic’s Claude Mythos — presented via Project Glasswing — is being billed as a leap in coding, reasoning and cybersecurity capabilities, and the company’s enterprise traction is growing fast. It has been reported that Anthropic’s annual recurring revenue (ARR) recently surpassed $30 billion, reportedly exceeding OpenAI’s ARR of about $25 billion; different accounting methods make direct comparisons imprecise, but the trend underscores a shift toward high enterprise token consumption.

Market pressure and strategic stakes

Why the interest in a guarded deployment? Because the next major model release is no longer solely a technical milestone; it is a commercial and narrative one. OpenAI faces investor timelines tied to a possible IPO window in the second half of 2026 and reportedly completed large rounds of financing, while rivals are touting lower costs, longer contexts and deeper workflow integration. If GPT‑6 or equivalents land as promised, they will need to prove not only superior performance but predictable revenue generation and enterprise lock‑in.

Security and geopolitical overtones

Any suggestion of a guarded team led by someone of Chinese descent will attract scrutiny given current geopolitical sensitivities: export controls, talent‑movement scrutiny and US‑China tensions have made personnel links a flashpoint in AI policy debates. Reportedly guarded or secretive security practices raise legitimate questions about intellectual property protection and operational transparency, but reportage remains unconfirmed. Who is safeguarding the “Glass Wing,” and why, remains an open question — one that observers of the global AI competition will watch closely.