Use use cache: remote to store cache entries in a remote handler (e.g. Redis or KV), providing durable, shared caching across all server instances.
Regular use cache stores entries in-memory, which works for static shell content but may have low hit rates at runtime in serverless environments.
Most useful for data that would otherwise be re-fetched on every request, where remote caching reduces load on rate-limited APIs, slow backends, or expensive operations.