news

Coinbase halves its AI bill with cheaper defaults, routing, and caching

Coinbase CEO Brian Armstrong says five levers — cheaper model defaults (GLM 5.2, Kimi 2.7), task routing, caching, lean context, and spend visibility — cut the company’s AI bill roughly in half despite rising token volume.

Published 2026-06-29Source: The Decoder
Generated Tokenmaxxing editorial thumbnail for Coinbase halves its AI bill with cheaper defaults, routing, and caching

Why it matters

It is a named-company playbook for cutting AI cost ~50% without rationing engineers: 91% had never bumped into the prior usage caps, so swapping defaults beat restricting access.

Tokenmaxxing read

Caching was the biggest lever, lifting cache hits from 5% to 60%; defaulting to open-weight Chinese models also squeezes Western labs on price right as several eye public listings. Fix the gateway before you ration people.

Source takeaway

Armstrong’s own account, relayed by The Decoder; the levers are directional, not audited line items, and the headline “half” covers spend while token volume still grows — keep those two separate when modeling savings.

Topic links

Related projects

Tools that match this angle

#1Direct
Routing

LiteLLM

BerriAI/litellm

An OpenAI-compatible gateway and SDK for calling many model providers with budgets, logging, load balancing, guardrails, and cost tracking.

63.6K13.4KSource-available
gatewaycost-trackingrouting
#2Direct
Observability

Langfuse

langfuse/langfuse

Open-source LLM engineering platform for observability, traces, metrics, evals, prompt management, datasets, and playground workflows.

37.7K4.3KSource-available
tracesevalscosts
#10Direct
Routing

Portkey Gateway

Portkey-AI/gateway

An AI gateway for routing across LLMs with guardrails, provider abstraction, and an OpenAI-compatible API surface.

13.4K1.4KMIT
gatewayguardrailsrouting
Related feed

More source-linked context

Generated Tokenmaxxing editorial thumbnail for The Hidden Cost of LLM Routing: Why Switching Models Costs More
long-formTD
long-form

The Hidden Cost of LLM Routing: Why Switching Models Costs More

Nhu Hoang replayed one 20-turn agent session under three routing policies on a vLLM server with prefix caching, then priced the token flows at Claude API rates. Switching models every turn saved 13%; the price sheet promised 40%.

tokenmaxxingmodel-routingmodel-router
Read note
Generated Tokenmaxxing editorial thumbnail for Enterprise AI budgets break at the handoff to production
long-formEC
long-form

Enterprise AI budgets break at the handoff to production

Express Computer interviews New Relic India's Ganesh Narasimhadevara on why AI bills keep climbing while the blended cost per million tokens has fallen over a year, from $18.40 in Q1 2025 down to $6.07 in Q1 2026.

tokenmaxxingai-spendcost-governance
Read note
Google Cloud Blog source artwork
newsGC
news

What sports cars can teach us about optimizing AI token spend

In an Aug. 26 Q&A, Mike Clark, a product director for Google's Gemini Enterprise Agent Platform, argues against tokenmaxxing and for what he calls efficiencymaxxing: the best business outcome per dollar, not the most tokens consumed.

tokenmaxxingagentstoken-consumption
Read note