# prompt-caching

Prompt caching should save you 90% on token costs — but one obscure bug can silently make you pay 10x more. From DANGEROUS_uncachedSystemPromptSection to the cch=00000 billing trap hidden in Claude Code's DRM, here's why prompt engineers now need to be accountants too.

Anthropic's prompt caching got major updates in 2026: Automatic Caching removes manual breakpoint headaches, 1-hour TTL keeps caches alive longer, and the invalidation hierarchy decides what blows up when you change things. Plus our real-world $13.86 billing disaster story.

Anthropic engineer Thariq shares Claude Code prompt-caching lessons: system prompt order matters, tools cannot change mid-conversation, switching models costs more than staying, and compaction must share the parent's prefix. Real SEV alerts included.

The 'Context Tax' in AI brings triple penalties: cost, latency, & reduced intelligence. Nicolas Bustamante's 13 Fintool techniques cut agent token bills by up to 90%. A real-money guide for optimizing AI context, covering KV cache, append-only context, & 200K token pricing.

An AI engineer stuffed user-specific data into the system prompt, watched his bill double, and learned his lesson. Plus six practical tips to consistently hit prompt cache. (Part 1 of 3)

Part 1 taught you how to save money. Part 2 explains why those tricks work. From the two stages of LLM inference (prefill/decode) to KV cache fundamentals to the GPU memory crisis that makes naive caching fall apart at scale. (Part 2 of 3)

Operating systems solved memory fragmentation with paging decades ago. vLLM brought that same trick to GPUs, added block hashing and prefix caching, and made prompt caching a reality. Series finale — every puzzle piece clicks into place.