Anthropic releases report stating Chinese AI companies including DeepSeek and Moonshot used third-party API proxies to bypass restrictions and access Claude

Date2026.09.10

LocationUnited States

Fact statusReported

Importance IV ●●●●●

Verification Reported · 2 Sources

On September 10, Anthropic released a threat intelligence report stating that Chinese AI companies, including DeepSeek and Moonshot, used third-party API proxy services to bypass regional restrictions and call Claude. The report also noted that these relay services may obtain model access through methods such as fake accounts and stolen API keys.

Evidence EVIDENCE

Sources 2 · Archived 0

SourceTypeReliabilityLink

李老师不是你老师

2026-10-24 06:17 · 网信办通报江苏某科技公司运营AI API中转站，因未开展安全评估及存在内容安全风险被警告处罚。

First-hand ●●●●●

Original

美国之音中文网

2026-10-24 06:30 · 美国之音中文网发布了新视频：美国之音中文广播 (2026年9月15日)

Media ●●●●●

Original

Context CONTEXT

Cause Not recorded yet

This event Anthropic releases report stating Chinese AI companies including DeepSeek and Moonshot used third-party API proxies to bypass restrictions and access Claude 2026.09.10 · Reported

Effect

Follow-up Anthropic names Zhipu and Moonshot again, claims to have banned tens of thousands of distillation accounts

10.12 ·Reported ·Importance III ·United States

Coverage of this event COVERAGE

Sayings · 2026-09-16 22:00 The Cyberspace Administration of China announced that a tech company in Jiangsu operating an AI API relay station was warned and penalized for failing to conduct security assessments and posing content security risks.

Article · 2026-09-14 22:00 Voice of America Chinese Radio Broadcast (September 15, 2026)

Report