The invisible software brain inside the world's biggest internet providers - turning raw network data into security, intelligence and calmer connected homes.
CUJO AI builds cloud-native software that broadband and mobile operators run inside their own networks. Instead of asking every household to install an app on every laptop, phone and smart bulb, the platform watches traffic at the network level, learns what each connected device is by how it behaves, and blocks threats like malware and phishing before they ever reach a screen.
The pitch to operators is simple: your network already produces the richest device data in technology, and most of it goes unused. CUJO AI turns that exhaust into three things carriers can sell or rely on - cybersecurity, device and network intelligence, and personalized subscriber experiences such as parental controls and privacy protection.
The name is a nod to a guard dog, and that is roughly the job: sit quietly at the edge of the home network and stop the bad things from getting in. The difference is scale. CUJO AI's models run across tens of millions of homes at once, and the company says the platform blocks more than 17,000 threats a minute.
Crucially, it is invisible by design. Subscribers rarely see the CUJO AI brand - they see a "security" toggle or a parental-controls screen inside their internet provider's app. The intelligence underneath is CUJO AI's.
"CUJO AI transforms raw network data into actionable intelligence - unlocking new revenue, improving operations and driving deeper user engagement at scale.— CUJO AI, company positioning
Network-based defense combining DNS, endpoint and IoT security to block malware, phishing and intrusions - no extra hardware for the subscriber.
Since 2018Machine learning identifies and classifies every connected device by how it behaves on the network, even ones the operator can't name.
Since 2016Near real-time analytics on device usage, application trends and threats - turning network data into operator insight and revenue.
Since 2019Privacy protection that limits online tracking and data leakage for the devices across a household.
Since 2020Profile-based content controls and digital life protection operators can offer as premium subscriber features.
Since 2018From the 2025 Domos acquisition: application-level Wi-Fi optimization focused on real outcomes - lag-free gaming, smooth video calls.
Since 2025CUJO AI sells to broadband and mobile operators, not to consumers. Its customers are among the largest carriers in North America and Europe, and in several cases they are also investors and deployment partners. Charter Communications, which led the Series B, went on to protect more than 24 million Spectrum subscribers with the technology.
Most cybersecurity companies chase the endpoint - the laptop, the phone, the app. CUJO AI went the other way and bet on the network layer, where the operator already sits and where a smart thermostat or camera that will never run security software still passes its traffic. That contrarian choice is the core of its differentiation: protection that scales without a per-device install, and privacy-respecting detection built for carrier scale.
It competes with the likes of Plume, Allot, F-Secure/WithSecure, Bitdefender and Akamai in the operator-security market. CUJO AI's angle is a single cloud-native platform that bundles security, device intelligence, analytics and, since 2025, Wi-Fi experience - reducing the number of vendors an operator has to stitch together.
CUJO AI began in 2015 as a consumer gadget - a firewall in a box that Indiegogo backers funded to 763% of goal, raising about $228,000. It was a hit, but the founders made a hard call in 2018: the box wasn't the business, the network was. The company shifted entirely to software-as-a-service for operators.
Today the model is recurring B2B revenue tied to household and device coverage. Carriers license the platform and bundle the features as value-added services, which turns security and analytics from a cost center into a product line. Third-party estimates put annual revenue in the region of $125 million, though the company does not publish figures.
The expertise is machine learning applied to network and device behavior - the kind of work that only pays off at scale and with constant iteration, which is why roughly 65% of the company is dedicated to research and development. Teams are spread across the US, UK, Finland, Lithuania, the Netherlands and Hungary.
That research focus earned outside recognition too: in 2018 the World Economic Forum named CUJO AI a Technology Pioneer, a list whose past honorees include Airbnb, Google and Spotify.
Einaras von Gravrock, Yuri Frayman and Saulius Tvarijonas launch CUJO AI; the smart-home device raises ~$228K on Indiegogo at 763% of goal.
The device protects around 3,200 homes, with a stated goal of one million by 2020.
Shifts to operator software, closes a strategic Series B led by Charter Communications, and is named a WEF Technology Pioneer.
Deploys with Charter to 24M+ Spectrum subscribers, launches CUJO AI Lens, and passes 500 million protected devices.
Introduces CUJO AI Incognito privacy protection and wins a gold Edison Award for innovation.
Acquires Norway's Domos to add Quality of Outcome Wi-Fi experience optimization across 60M+ homes.
It provides cloud-native software that lets broadband and mobile operators identify connected devices, block cyber threats in real time, and deliver security, parental controls, privacy and network analytics to their subscribers.
Not anymore. CUJO AI started as a consumer hardware device but pivoted in 2018 to selling software to network operators, so most people use it invisibly through their internet provider.
Major broadband and mobile operators including Comcast, Charter, TELUS, Rogers, Cox, Shaw, Sky, BT/EE, Deutsche Telekom and T-Mobile US, covering 60+ million homes.
It was founded in 2015 by Einaras von Gravrock, Yuri Frayman and Saulius Tvarijonas.
It works at the network level, using machine learning to analyze device behavior and traffic so it can classify devices and block threats before they reach endpoints - no per-device app required.