Great reporting by Juha Saarinen at iTnews on the residential proxy ecosystem fueled by AI data demand.
Proud to see IPinfo and Ben Dowling contributing data and analysis to the story. The key finding: brand-level takedowns like the GTIG/IPIDEA disruption don’t remove devices from circulation. IPinfo’s data showed 74–88% IP overlap across providers, with 46% of residential proxy IPs appearing in multiple provider pools simultaneously. In 90 days across just six tracked providers, nearly 79 million unique IPs were identified globally.
That’s the reality of this ecosystem. The infrastructure is shared, resilient, and deeply interconnected. Take down one brand, and the same devices keep routing traffic through sister providers under different names. Five of the providers named by Google shared the same backend network operator — pointing to common control behind multiple storefronts.
This is exactly the kind of problem that evidence-based IP intelligence is built to address. Static blocklists and single-signal detection fall short when proxy networks rotate IPs rapidly across services. Effective detection requires combining active measurement, direct verification through proxy services, and temporal signals like recency and persistence — so you can distinguish between an IP that’s currently part of a proxy pool and one that’s returned to normal residential use.
The article also highlights an important consumer angle. Most users who install bandwidth-sharing apps have no idea their device becomes an exit node for third-party traffic. Their IP reputation degrades, their home network gets exposed, and in cases like the Kimwolf botnet, attackers can tunnel back through the proxy infrastructure to reach local devices.
Residential proxy traffic is now a structural feature of the internet, not an edge case. Glad IPinfo’s data could help make the scale of this visible.