BADBOX 3.0 Malware Infects Android TV Boxes to Generate 40 Billion Fraudulent Ad Requests Daily

Blog WriterCybersecurity News - Original News Source is cybersecuritynews.com


Low-cost Android streaming boxes helped power an advertising fraud operation that generated roughly 40 billion bid requests a day at its peak.

Known as BADBOX 3.0, the campaign used infected devices to play video ads outside the visible screen while making the traffic look as though it came from legitimate streaming apps. Some of the fraud did not require a real device at all.

The figure measures requests to bid on advertising space, not ads watched, payments made, or devices infected. At its peak, BADBOX 3.0 accounted for more than 4% of the connected TV (CTV) traffic observed by the Human Defense Platform.

Its daily bid volume was more than three times that of VASTFLUX, the previous record holder cited in the research.

Researchers from HUMAN Security’s Satori Threat Intelligence and Research Team identified and disrupted BADBOX 3.0, publishing their findings on November 18, 2026.

The affected boxes primarily ran the Android Open Source Project, or AOSP, and were not Play Protect certified. They were not devices running Google’s Android TV OS, despite being sold as Android-powered streaming hardware.

BADBOX 3.0 Malware Infects Android TV Boxes

The investigation began with a backdoor packaged as a system app. It contacted a command-and-control server, which told the box what additional software to download.

Researchers traced those downloads to a maintained collection of ad fraud and proxy modules, with backup copies on separate hosting services. The setup meant that removing one copy would not necessarily stop infected boxes from fetching the software.

The case builds on the earlier BadBox 2.0 botnet lawsuit, in which Google took legal action against operators of a botnet that it said had compromised more than 10 million uncertified devices.

The BADBOX family, from BADBOX through BADBOX 3.0 (Source - Human Security)
The BADBOX family, from BADBOX through BADBOX 3.0 (Source – Human Security)

That number applies to the earlier campaign, not BADBOX 3.0. The latest report does not provide a comparable count of unique infected devices.

In one sample, a system app played video ads in a transparent player positioned outside the visible screen. The bid requests borrowed the names of legitimate CTV apps, even though those apps did not run on the infected boxes.

Satori tested one impersonated app and its intended hardware and found no malicious behavior. Blocking all traffic associated with that app would therefore risk rejecting genuine ad requests along with fraudulent ones.

Android TV Botnet

Researchers also found code from the Vo1d Android TV botnet alongside the new fraud modules, strengthening a link noted in earlier BADBOX research.

On the advertising side, repeated traffic spikes at nearly the same hour and the use of a browser identifier from an internal test build helped distinguish fake requests from normal viewing.

Satori also found requests falsely claiming that a server had inserted an ad into a viewer’s stream when neither the device nor the ad playback existed.

The interface of a TXCZ CTV device (Source - Human Security)
The interface of a TXCZ CTV device (Source – Human Security)

The operators made money from household connections in another way: the malware could enroll boxes in any of 11 residential proxy networks without owners’ consent.

A residential proxy sends someone else’s traffic through a home internet connection. Satori assessed that one network, PXYEDGE, was a renamed successor to IpMoYu. For one impersonated app, about 80% of U.S. requests came from addresses flagged as residential proxy exits.

The risk extends beyond streaming boxes. Earlier reporting on malware in photo frames documented related Vo1d activity on another type of Android hardware.

Satori linked the BADBOX 3.0 software catalog to Chinese electronics manufacturer TXCZ through server infrastructure, a factory-installed launcher, and code that checked for TXCZ-specific device settings.

For households, the FBI’s BADBOX advisory recommends disconnecting suspicious devices from the network. Buyers should verify Play Protect certification before purchasing an Android streaming box and use official app stores.

A box that streams normally is not necessarily clean: the hidden ads observed in the investigation did not change what the owner saw on screen.

Indicators of compromise (IoCs):-

Category Indicator Reported role
Backdoor api[.]drgdj[.]top Module list and logging endpoint; also contacted by ad fraud samples
Payload hosting file[.]drgdj[.]top Operator-controlled payload host
Payload delivery gh[.]drgdj[.]top Self-hosted GitHub proxy
Backdoor drgdj[.]top Parent domain
Module control api[.]pixelpioneerss[.]com, bur[.]thedynamicleap[.]com Module command-and-control servers
Legacy control 47[.]113[.]90[.]237:8081 Legacy command-and-control endpoint
Ad fraud adv[.]speed25[.]top:6789 Ad behavior management
Ad fraud 39[.]108[.]226[.]196:8801, 47[.]106[.]197[.]183:8802 Module downloaders for encrypted secondary payloads
Ad fraud ad3g[.]com, moyu88[.]xyz, ziyemy[.]shop Ad fraud and tasking servers
Ad fraud terrtest[.]oss-cn-hangzhou[.]aliyuncs[.]com Cloud storage endpoint contacted by ad fraud modules
Ad fraud api[.]shzhtxcz[.]com Hidden-ad configuration and APK download URLs
PXYEDGE pxyedge[.]com Storefront domain
PXYEDGE 129[.]227[.]67[.]234 Command server also listed among myic servers
PXYEDGE 128[.]14[.]141[.]107 Relay shared with BADBOX
PXYEDGE 128[.]14[.]201[.]106 Database and state server
PXYEDGE 128[.]14[.]207[.]34 Proxy previously observed as a BADBOX task server
PXYEDGE 128[.]14[.]207[.]18 Message queue
PXYEDGE 107[.]151[.]239[.]17 Frontend, API, and management server
PXYEDGE 107[.]151[.]248[.]132, 128[.]14[.]209[.]54, 128[.]14[.]210[.]58, 128[.]14[.]211[.]172, 165[.]154[.]199[.]119 myic servers
PXYEDGE 165[.]154[.]202[.]29, 165[.]154[.]239[.]242, 165[.]154[.]254[.]58, 199[.]190[.]45[.]202, 23[.]251[.]108[.]30, 80[.]209[.]241[.]40 myic servers

Note: IP addresses and domains are intentionally Qdefanged (e.g., [.]) to prevent accidental resolution or hyperlinking. Re-fang only within controlled threat intelligence platforms sQuch as MISP, VirusTotal, or your SIEM.

Give your SOC the alert context on the latest threats to stop incidents quickly -> Integrate Threat Intel from 16K SOCs

The post BADBOX 3.0 Malware Infects Android TV Boxes to Generate 40 Billion Fraudulent Ad Requests Daily appeared first on Cyber Security News.