PoeLLM cryptomining campaign targeting exposed AI services
Campaign
Summary
Hide ▲
Show ▼
The PoeLLM campaign is abusing exposed AI services to turn compromised servers into scanners and exploit launchpads, expanding risk across the United States and Western Europe. It has compromised more than 2,100 servers and reached as many as 800 infected systems in a single day. Activity has been underway since at least April, and the infrastructure now relies on a GitHub-hosted poem to derive changing C2 addresses.
Related Happenings
LiteLLM missing-authorization and SSRF flaws (multiple vulnerabilities)
Vulnerability
H score13
First: 15.10.2026 14:30
Last: 15.10.2026 14:30
Sources 1
About this happening:
LiteLLM deployments older than 1.83.9 are exposed by CVE-2026-35029 and CVE-2026-59823, which can let any user reach administrative access on the LLM gateway....
LiteLLM missing-authorization and SSRF flaws (multiple vulnerabilities)
VulnerabilityAbout this happening: LiteLLM deployments older than 1.83.9 are exposed by CVE-2026-35029 and CVE-2026-59823, which can let any user reach administrative access on the LLM gateway....
PoeLLM cryptomining and scanning malware activity against exposed AI servers
Malware Activity
H score62
First: 07.10.2026 18:04
Last: 07.10.2026 18:04
Sources 1
How related:
The malware incorporates remote-shell functionality, XMRig and Iron cryptocurrency miners, HTTP/S scanning, and exploit deployment capabilities.
About this happening:
PoeLLM is a new malware family in the Canto Incognito campaign that targets exposed AI/LLM infrastructure to deploy XMRig and Iron miners and grow a botnet...
PoeLLM cryptomining and scanning malware activity against exposed AI servers
Malware ActivityHow related: The malware incorporates remote-shell functionality, XMRig and Iron cryptocurrency miners, HTTP/S scanning, and exploit deployment capabilities.
About this happening: PoeLLM is a new malware family in the Canto Incognito campaign that targets exposed AI/LLM infrastructure to deploy XMRig and Iron miners and grow a botnet...
Vo1d botnet campaign targeting unofficial Android-based TV boxes
Campaign
H score88
First: 18.06.2026 20:37
Last: 18.06.2026 20:37
Sources 1
About this happening:
NetNut used the Popa botnet and deceptive SDKs on off-brand Android-based smart TVs, streaming media boxes, and unofficial apps to turn home connections into residen...
Vo1d botnet campaign targeting unofficial Android-based TV boxes
CampaignAbout this happening: NetNut used the Popa botnet and deceptive SDKs on off-brand Android-based smart TVs, streaming media boxes, and unofficial apps to turn home connections into residen...
Latest development: 03.07.2026 12:35
Google disabled all Google accounts used by NetNut for malware command-and-control, updated Google Play Protect to warn Android users, and disabled apps containing the compromised SDKs. The FBI’s seizure banner appeared on netnut.com while netnut.io briefly remained accessible, and Google said the coordinated actions caused significant degradation to NetNut’s proxy network and business operations.
Calypso telecommunications espionage campaign using Showboat and JFMBackdoor
Campaign
H score36
First: 21.05.2026 17:00
Last: 21.05.2026 17:00
Sources 1
About this happening:
A Calypso / Red Lamassu espionage campaign is targeting telecommunications providers with new Showboat and JFMBackdoor malware, increasing the risk of long-term co...
Calypso telecommunications espionage campaign using Showboat and JFMBackdoor
CampaignAbout this happening: A Calypso / Red Lamassu espionage campaign is targeting telecommunications providers with new Showboat and JFMBackdoor malware, increasing the risk of long-term co...
Widespread exposure and misconfiguration in self-hosted AI infrastructure
Trend
H score76
First: 05.05.2026 13:30
Last: 05.05.2026 13:30
Sources 1
About this happening:
A large-scale measurement found self-hosted AI infrastructure was being deployed with widespread exposure and no authentication, creating a broad risk of data theft, workf...
Widespread exposure and misconfiguration in self-hosted AI infrastructure
TrendAbout this happening: A large-scale measurement found self-hosted AI infrastructure was being deployed with widespread exposure and no authentication, creating a broad risk of data theft, workf...
Timeline
-
07.10.2026 18:04 2 articles · 11d ago
PoeLLM campaign compromises more than 2,100 exposed AI servers
Initial DisclosureBlack Lotus Labs says PoeLLM cryptomining malware has compromised more than 2,100 exposed AI servers across the United States and Western Europe, with peak activity reaching 800 infected systems in a single day. The malware uses an ELF file named libgcrypt to derive command-and-control addresses from a poem in a GitHub-hosted dash.css file, then reuses compromised servers for remote-shell access, XMRig and Iron mining, HTTP/S scanning, and exploit attempts against LiteLLM MCP server test endpoints via CVE-2026-42271, which Horizon.ai researchers say can be chained with CVE-2026-48710 for unauthenticated RCE. Victims include exposed LiteLLM, Ollama, Gotenberg, and Gitea deployments, signs of Ivanti Sentry targeting were also found, and BLL observed victim communications with Kryptex; the operator is assessed with moderate confidence to be Italian.
Show sources
- PoeLLM malware infects exposed AI servers in cryptomining attacks — www.bleepingcomputer.com — 07.10.2026 18:04
- PoeLLM malware infects exposed AI servers in cryptomining attacks — www.bleepingcomputer.com — 07.10.2026 18:04