Tensormesh

Tensormesh is an AI inference platform founded by the creators of LMCache, focused on building a production-grade context caching layer for AI teams. The company helps developers reduce repeated model work, lower inference costs, and scale context-heavy workloads—such as agent workflows, RAG systems, and multi-turn conversations—by caching and reusing KV context across requests. Headquartered in Foster City, California, Tensormesh was co-founded by University of Chicago researchers who won the ACM EuroSys 2025 Best Paper Award for their CacheBlend research.