Sign InOpen Brain
AI EngineerVideoSource Linked

The Frontier AI Inference Cloud for Agents — Byung-Gon (Gon) Chun, FriendliAI

Agent-serving systems should optimize whole-task latency, preserving shared prefixes across tool calls and routing later steps toward the GPU that already holds their KV cache.

AI Engineer · Sep 19, 2026
Open Source Open MarkdownOpen JSON
Source Summary

Agent work alternates model calls with tools, grows its context, and often branches into subagents. Consecutive steps share large prefixes, so FriendliAI centers its stack on **prefix caching**, hierarchical and distributed KV storage, cache-aware routing, and agent-aware scheduling.

Practical Implication

Measure end-to-end task completion rather than treating every call as independent. Keep related steps cache-local, balance that locality against hot spots, and use task context when deciding preemption, eviction, or speculative prefill.

Agent-Ready Context
Agent work alternates model calls with tools, grows its context, and often branches into subagents. Consecutive steps share large prefixes, so FriendliAI centers its stack on **prefix caching**, hierarchical and distributed KV storage, cache-aware routing, and agent-aware scheduling.

Measure end-to-end task completion rather than treating every call as independent. Keep related steps cache-local, balance that locality against hot spots, and use task context when deciding preemption, eviction, or speculative prefill.

The talk’s speed, cost, and quality comparisons are provider-reported and do not include enough methodology here for independent evaluation. Agent-aware scheduling is also presented as an emerging optimization area, not a settled recipe for every workload.
Connected Context · Feed7 Judgment

This strengthens the case that agent inference should be optimized as a stateful, end-to-end workload rather than isolated requests. It extends prior cache-aware routing evidence with hierarchical KV storage and task-aware scheduling, while keeping the same qualification: locality must be balanced against load and validated on real traffic, not accepted from provider comparisons alone.

Vertical Mobility: Inference from MVP to Trillion-Parameter Workloads — Sitanshu Gupta, CoreWeaveCoreWeave independently supports the premise that repeated agent prefixes make KV-local routing a practical latency and cost lever.KV Cache-Aware Routing and P/D Disaggregation on Kubernetes — Yuchen Fama & Ashish Kamra, Red HatRed Hat supplies case-study evidence for cache-aware scheduling while clarifying that disaggregation depends on concurrency, prompt shape, and network performance.TokTier: Exact Stateful Tokenization for Agentic LLM ServingTokTier identifies repeated tokenization as a complementary bottleneck, showing that KV-cache locality alone does not guarantee low time to first token.Operating Distributed Inference Systems at Scale — Nishant Gupta & Naman Ahuja, MetaMeta reinforces the shift toward workflow-aware orchestration by likewise requiring scheduling to account for KV state, retries, hardware, and completed-task cost.
Context Map
infracoding#context-caching#cloud-agents#agent-reliability
Uncertainty
The talk’s speed, cost, and quality comparisons are provider-reported and do not include enough methodology here for independent evaluation. Agent-aware scheduling is also presented as an emerging optimization area, not a settled recipe for every workload.