IP Geolocation API latency

Our 'IP Lookup Tool - Geo-locate Any IP Address & Get Details' relies heavily on robust IP Geolocation API integration to provide its core functionality. We're currently facing significant challenges with our primary Geolocation API provider, specifically concerning inconsistent latency and data freshness. This is particularly noticeable during high-volume lookup periods and when querying IP addresses originating from specific, often more remote, geographic regions. We've attempted various optimizations, including implementing aggressive client-side caching strategies, deploying localized edge servers to reduce network hops, and even database sharding for pre-cached IP ranges. However, despite these efforts, we haven't been able to consistently resolve the fundamental trade-offs between achieving low latency and maintaining real-time data accuracy, especially given the dynamic nature of IP allocations and the varying data update cycles across different providers. What advanced strategies are fellow developers employing to achieve reliable sub-200ms global response times for IP Geolocation API calls while simultaneously maintaining real-time data accuracy, particularly when navigating the complexities of dynamic IP allocations and the diverse data update frequencies of multiple Geolocation API providers?