Case wiz-llmjacking-cloud-billing-adsLLMjackingWorldwide
Search-ad phishing hijacks cloud billing accounts for unauthorized LLM API usage
Wiz CIRT says attackers bought Google Search ads for terms like 'google cloud' that led to fake Google login pages, captured credentials and MFA codes, and then linked attacker-controlled GCP projects to victims' billing accounts. Within 24 to 48 hours victim organisations saw charges of tens to hundreds of thousands of dollars in unauthorized Gemini API usage. Google Ads removed the malicious campaign after it was reported.
HighAttack on AIStatus: Under investigationVerified case file
Open in the interactive mapCite this case
- Reported
- Oct 20, 2026
- Location
- Worldwide
- Target
- Organisations with privileged Google Cloud (GCP) users and billing accounts
- Severity
- High
- Verification
- Reported by credible press or researchers
- Triggered by
- Unknown threat actor(s)
- Actor type
- Criminal
- Sector
- Multiple / cross-sector
- Reported loss
- None reported
- CVE
- None
What happened
Wiz CIRT says attackers bought Google Search ads for terms like 'google cloud' that led to fake Google login pages, captured credentials and MFA codes, and then linked attacker-controlled GCP projects to victims' billing accounts. Within 24 to 48 hours victim organisations saw charges of tens to hundreds of thousands of dollars in unauthorized Gemini API usage. Google Ads removed the malicious campaign after it was reported.
Impact: Hundreds of thousands of dollars in unauthorized AI API usage were incurred, according to Wiz Research.
Who triggered it
Actor: Unknown threat actor(s)
Attribution: Wiz Research describes the threat actors but the summary does not name them.
Motive: Financial gain through free use of LLM APIs billed to victims.
How it happened
Attack vector: Malicious Google Search ad leading to phishing, credential theft, then hijack of the cloud billing account.
Root cause: Victims' credentials were phished through a malicious search ad, giving attackers control of billing accounts.
Timeline
- Reported by Wiz Research
Response
Wiz Research published its findings. Other responses are not stated in the summary.
Current status: Under investigation
Research warnings this case matches
- Seen in real incidents Privacy and security · MIT AI Risk Repository
Compared with published AI risk research from several institutes on the Research vs reality page.
Sources (open in a new tab)
- Wiz Research wiz.io
- Cloud Security Alliance Labs labs.cloudsecurityalliance.org
Cite this case
Search-ad phishing hijacks cloud billing accounts for unauthorized LLM API usage. AI Threat Watch, case wiz-llmjacking-cloud-billing-ads, Oct 20, 2026. https://aithreat.io/cases/wiz-llmjacking-cloud-billing-ads/
Permanent link: https://aithreat.io/cases/wiz-llmjacking-cloud-billing-ads/
Related cases
The same type of incident, most recent first.