---
format: "aidr-story-markdown/v1"
id: "dc936f74d4c8b0ef9e3ce4e8f8071e099fff108a1a2a2f771130d6c305f1853c"
canonical_url: "https://aidr.today/dc936f74?lang=vi"
title: "Annualized Cyber Incidents More Than Double to Nearly 10,000 Since Late Summer, Bridgewater Says"
lang: "en"
requested_lang: "vi"
available_langs: ["en"]
translation_fallback: "en"
fallback_fields: ["title","summary"]
published_at: "2026-11-19T20:52:36.000Z"
category: "Agents"
topics: ["safety","anthropic"]
source_urls: ["https://marketbrief.now/cybersecurity/annualized-cyber-incidents-more-than-double-to-nearly-10000-since-late-s-c47471d8","https://huggingnews.com/cybersecurity/annualized-cyber-incidents-more-than-double-to-nearly-10000-since-late-s-c47471d8"]
summary: "Bridgewater estimates that total cyber incidents have more than doubled since late summer, rising from approximately 4,000 a year to nearly 10,000 on a 60-day annualized basis. This spike follows a decade of gradual growth in which the count of incidents increased from about 1,500 in 2017 to roughly 4,000 by mid-2026. The surge coincided with the release of open weights for GLM-5.3, an AI model from China's Zhipu AI that Anthropic's red team found could build working cyber exploits. Verified AI-driven cyber incidents jumped from near zero to more than 200 a year in the same period, including campaigns in which AI agents stole bank and card data in Korea."
---

# Annualized Cyber Incidents More Than Double to Nearly 10,000 Since Late Summer, Bridgewater Says

> [Open the canonical story](<https://aidr.today/dc936f74?lang=vi>)

**Published:** 2026-11-19T20:52:36.000Z
**Category:** Agents
**Topics:** safety, anthropic

_English fallback used for: title, summary._

## Summary

Bridgewater estimates that total cyber incidents have more than doubled since late summer, rising from approximately 4,000 a year to nearly 10,000 on a 60\-day annualized basis\. This spike follows a decade of gradual growth in which the count of incidents increased from about 1,500 in 2017 to roughly 4,000 by mid\-2026\. The surge coincided with the release of open weights for GLM\-5\.3, an AI model from China's Zhipu AI that Anthropic's red team found could build working cyber exploits\. Verified AI\-driven cyber incidents jumped from near zero to more than 200 a year in the same period, including campaigns in which AI agents stole bank and card data in Korea\.

## Sources

- [Story source](<https://marketbrief.now/cybersecurity/annualized-cyber-incidents-more-than-double-to-nearly-10000-since-late-s-c47471d8>)
- [Story source](<https://huggingnews.com/cybersecurity/annualized-cyber-incidents-more-than-double-to-nearly-10000-since-late-s-c47471d8>)

