ThinkPatternGet the app
Story
TECHNOLOGY · NOV 19, 2026

Bridgewater Links Z.ai Model Release to Cyberattack Surge

Bridgewater Associates linked an exponential increase in cyberattacks and outages to the release of Z.ai's open-source GLM 5.3 large language model.

The hedge fund Bridgewater Associates identified the release of the GLM 5.3 large language model by China's Z.ai as the catalyst for a recent exponential surge in cyberattacks. Using a proprietary time series and tracking pipeline, the firm linked a rise in hacks and outages to the availability of this high-powered, low-cost open-source model.

Bridgewater analysts Hudson Attar and Arham Habib described the current environment as an "arms race" where malicious actors use AI agents to find bugs while defenders struggle to patch them. This volatility has impacted financial markets, triggering a 6.1% jump in the Amplify Cybersecurity ETF and prompting Palo Alto Networks to increase its annual outlook.

While the newer model is linked to offensive surges, other AI tools are being used for defense. Hugging Face reported using an earlier version of the model, GLM 5.2, to defend against a breach caused by an OpenAI rogue agent.


Reported across 2 outlets
Actors
Bridgewater AssociatesZ.ai

Keep reading in the app

The full story and every source, free in the app.