Xbow researchers used AI to find Chrome exploit chain, earning $250K bounty

Researchers at Xbow USA identified a chain of vulnerabilities in Google’s Chrome browser, an effort that earned them a $250,000 bug bounty from the vendor. This reward marks the first payout under Google’s new 2026 rules, according to Computerworld. The discovery involved using artificial intelligence tools to accelerate the research process.

The security team spent just six days developing the exploit chain, incurring approximately $40,000 in AI compute costs. They utilized OpenAI’s GPT-5.6 Sol model to demonstrate the potential of publicly available AI systems in security research. The researchers maintained full control over the strategy, deciding where to look in the code and verifying the AI’s output, rather than relying on the model for autonomous decision-making.

Technical details and remediation

The vulnerability chain allowed an attacker to escape the sandbox containing Chrome’s V8 JavaScript engine. Following that initial escape, the exploit could break out of the operating-system process sandbox as well. This sequence would enable a malicious web page to launch a Windows application.

Xbow USA presented these findings to Google, which subsequently fixed the flaws. The researchers published a detailed blog post outlining how they developed the exploit. As Threat Times reported on October 16, 2026, Google regularly releases updates to address security issues in its browsers.

Scope and impact

The researchers noted that this specific chain of vulnerabilities could have enabled the compromise of one billion Windows machines worldwide. The exploit targeted the Windows version of the browser, leveraging the sandbox escape capabilities to gain control over the host system.

By using a general-purpose AI model rather than a specialized security implementation, the team aimed to show what is possible with accessible technology. The financial cost of the compute time was minimal compared to the potential impact of the vulnerability.

What should you do?

Organizations should ensure that all endpoints running Google Chrome are updated to the latest version provided by the vendor. Since Google has fixed the flaws, applying the patch is the primary mitigation step. Administrators should monitor for any further disclosures related to these specific components.

What to do and how to stay safe: Google Chrome

• Update all Windows machines running Chrome to the latest stable version immediately to apply the vendor's fix.
• Verify that automated update mechanisms are functioning correctly to ensure future security patches are installed without delay.
• Review browser extension policies to limit the attack surface, as malicious extensions can sometimes bypass sandbox protections.
• Monitor official security advisories from Google for any additional details regarding these vulnerabilities or related issues.

Step-by-step guide: Mobile Device Security: How to Protect Data Beyond Passwords

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

How much did it cost the researchers to find this vulnerability?

The researchers spent approximately $40,000 on AI compute time during the six-day investigation.

Did the AI model find the vulnerability independently?

No, the researchers set the strategy, chose where to look, and checked the AI's work, using it primarily to search code and build proofs of concept.

Has Google fixed the vulnerabilities?

Yes, Xbow USA presented the findings to Google, and the vendor has fixed the flaws.