Gravity matters. I caught up with a friend today, and we talked about Mooncake
@KVCache_AI â by 2026, whether for PD disaggregation or KV storage, Mooncake has become a first-class citizen and the default solution across the top 3 LLM inference engines.