Google’s Threat Intelligence Group says the XCPE botnet reached more than 1.3 million Android TV devices. It also found infections in other Android connected equipment, including telehealth devices, karaoke machines, and digital signage. The case raises a broader issue about trust after setup.

OpenAI, Micro1 and Anthropic have paid out billions of dollars to partner with companies on their operational data. It's creating a new revenue stream for businesses over 20 people and pays anywhere from $100k - $2 million a pop.

Get an estimate here

Google says the infected devices came from Videostrong, a manufacturer of certified Android hardware. Certification gives a device access to Google applications and services, according to Ars Technica’s account of the process. The incident shows that a check before approval cannot settle every question about software behavior after setup.

The reported sequence matters. Google describes an initial component that contacted its operators and enabled later malware modules. Ars reports that the backdoor became active after devices were online, then installed additional software with system-level permissions. Google’s technical report says the malware could place files in Android’s system partition and request privileges for installing packages and restarting devices.

That design changes the security boundary. A device may pass initial review, while later network activity supplies code that was not visible during approval. This does not show that every certified product is unsafe.

The next wave of millionaires will be people who figured out how to make AI work for them.

The window to get ahead is still open. But not for long.

Here are 200+ proven ways to make money with AI in 2026.

Sign up for Superhuman AI, the free daily newsletter read by 1M+ professionals, and get instant access to all 200+ ways to profit from AI this year.

Claim your free list

The botnet’s functions also show why compromised consumer hardware matters beyond its owner. Google observed modules for residential proxy traffic, bandwidth sharing, and advertising fraud. Ars reports that investigators also saw operators able to run commands, capture screens, and inspect devices on local networks. These functions can support activity beyond the device owner.

The scale figures require attribution. Google reported more than 1.3 million affected devices. Trend Micro counted 1,472,554, according to Ars, which describes that number as somewhat higher. The Record reported infections in more than 200 countries. These figures come from different investigators and should not be collapsed into one exact total.

Google says it took legal action against domains and servers used to control the botnet, and ended its relationship with Videostrong. It also says it barred the company and affiliates from future certified-device work. Those steps disrupt operator access and constrain one manufacturer’s route into the certified ecosystem.

Disruption does not automatically mean every device is clean. Ars reports that some malware modules remained installed after the command infrastructure was cut off, and Play Protect had been disabled. The Record likewise says investigators did not describe a customer notification or forced cleanup effort. The evidence distinguishes network disruption from verified repair.

That distinction is central to incident response. Taking down command servers can reduce an attacker’s ability to issue new instructions. It does not, by itself, demonstrate that installed software has been removed, security settings restored, or devices returned to their original state. A reliable closure report would need device-level remediation evidence.

For manufacturers and platform operators, the case points to several separate checkpoints. Build review must examine what ships in firmware. Certification should consider behavior after a device connects to the internet. Update systems need a way to correct persistent components. Monitoring must identify unexpected permissions, hidden services, and changes that arrive after approval.

For buyers, the public reporting is not enough to identify every affected unit or prescribe one universal remedy. Google has not publicly explained a notification process in the sources reviewed here. A model list can guide checks, but cannot establish an individual device’s condition.

The larger lesson is precise. Certification can reduce risk, yet it cannot replace ongoing review of software, updates, and device behavior. This case file is one check-in, not proof every endpoint has been repaired. Use it to ask what validation covers after setup.

Instead of running your deals across scattered email threads and forgotten attachments, Aligned brings everything into one shared deal workspace.

Buyers get a single place to evaluate, loop in their team, and say yes. Sellers get visibility into what is actually happening when they're not in the room and AI Deal Insights that help you act accordingly, before it's too late.

Best part? Your first deal room takes only minutes to stand up.

Start Using Aligned Now