IAB Report: Privacy Proxies Are Challenging Traditional IP Geolocation

An Internet Architecture Board workshop report published in July 2026 says the assumptions behind traditional IP geolocation are being stretched by privacy proxies, satellite networks, new deployment models, and rising privacy expectations. The report does not declare IP geolocation obsolete. Instead, it describes a system under pressure: applications still need coarse regional context, while an IP address is becoming a less dependable passive identifier for a person’s actual location.

What the report highlights

The workshop brought together network operators, application providers, geolocation specialists, privacy experts, and policy stakeholders. A central observation is that IP addresses were designed for routing, not for locating end users. Databases infer location from registry data, routing, operator submissions, commercial observations, and other signals.

New network architectures make those inferences harder. Satellite networks can move traffic through gateways far from the user. Privacy proxies deliberately separate the client address from the address seen by a destination. Mobile and cloud networks can relocate prefixes or centralize egress. Each case can create a gap between network location and user location.

Why privacy proxies matter

Privacy proxies reduce passive linkability by preventing destinations from seeing the original client address. This improves privacy, but it also weakens applications that silently treat the observed address as a precise location signal.

The report frames this as a design trade-off rather than a proxy defect. A privacy-preserving system should not be expected to reveal more about a user than its architecture intends. Applications that truly need location may require consent-based application signals, account settings, or other transparent mechanisms.

The practical effect on proxy evaluation

For proxy buyers, the report reinforces the need to separate three questions:

1. Where is the proxy exit network announced and operated?
2. How do major location datasets classify that exit?
3. What location does the target application infer after combining IP, account, language, device, and session signals?

These answers may differ. A provider can operate an exit in the promised country while a stale database reports an older prefix location. Conversely, a database match alone does not prove physical proximity or residential network characteristics.

Country confidence is different from city precision

Country-level classification is usually more stable than city-level inference. The workshop report describes increasing tension between coarse location needs and precise user tracking. Operational testing should therefore define the minimum useful granularity.

For localization or market research, country or region may be sufficient. City-level claims should be tested across multiple datasets, ASN information, latency, DNS behavior, and repeated sessions. Results should be expressed as confidence, not an exact coordinate.

What applications may change

The report notes growing interest in consensual application-layer location mechanisms. These can provide a user-approved location when an application genuinely needs it, instead of extracting precise meaning from a passive network identifier.

For services, this may lead to a layered model: IP for coarse routing and abuse controls, explicit signals for user-facing location, and clear fallbacks when signals disagree. Privacy proxies make those boundaries more visible.

Operational recommendations

• Test country, region, city, ASN, DNS, and session stability separately.
• Record the database and observation time for every location result.
• Expect temporary disagreements after prefix moves or routing changes.
• Do not use latency as proof of exact physical location.
• Treat target-site classifications as target-specific evidence.
• Avoid collecting more precise location data than the workflow requires.
• Provide a correction path for prefixes that are consistently misclassified.

What this means for 98IP users

Regional proxy testing should use a documented acceptance matrix rather than a single lookup result. Buyers should decide whether the workflow requires country-level routing, city-level confidence, a particular network type, or stable sessions. Those requirements determine which failures matter.

For authorized regional testing, available locations and session models are listed on the 98IP English site. Always follow target terms, applicable law, and data-minimization principles.

Source note: Internet Architecture Board, Report from the IAB Workshop on IP Address Geolocation, July 2026. The report remains an Internet-Draft and may change before final publication.