CG-NAT
CG-NAT (Carrier-Grade Network Address Translation) is a technique where a mobile or ISP carrier shares one public IPv4 address across many subscribers. It's why mobile proxy IPs often represent thousands of cellular users at once.
Definition
CG-NAT (Carrier-Grade Network Address Translation) is a technique where a mobile or ISP carrier shares one public IPv4 address across many subscribers by NAT-translating their private addresses to the shared public one. A single CG-NAT public IP often represents thousands of cellular users at once.
The practice is ubiquitous on mobile networks (where IPv4 address scarcity and subscriber density make it necessary) and increasingly common on fixed-line ISPs that have exhausted IPv4 pools.
Why CG-NAT matters for AI workloads
CG-NAT has two specific implications for proxy routing in AI contexts:
-
Mobile proxy traffic looks like thousands of users sharing an IP. Anti-bot systems know this and set their thresholds accordingly. A single mobile IP can legitimately serve thousands of requests per day from distinct human users; the threshold for "suspicious" is much higher than on fixed residential.
-
Fingerprint + behaviour matter more than IP reputation. Because CG-NAT IPs are shared, the IP itself is a weaker signal. Anti-bot systems compensate by weighting browser fingerprint, request timing, and behaviour patterns more heavily. A mobile-origin AI request with a clean fingerprint is usually fine; one with a datacenter-like fingerprint is suspicious despite the mobile IP.
CG-NAT and cellular-anchored AI eval
For evaluation workloads that measure how a model API treats cellular origin vs. fixed-line origin, CG-NAT is the specific network layer that makes cellular-origin traffic distinct. See our 4G mobile proxy page for the pool specifics and residential vs datacenter routing matrix for the broader class tradeoffs.
Related
Ship on a proxy network you can actually call your ops team about
Real ASNs, real edge capacity, and an engineer who answers your Slack the first time.