Blockchain exchange Coinbase announced that it has shifted its core business operations to use more cost-effective Chinese AI models, such as GLM5.2 and Kimi2.7. This move has significantly increased token usage while successfully cutting the company's AI costs in half. Coinbase CEO Brian Armstrong revealed that although developers still retain the right to choose models, the newly established automatic routing system can intelligently distribute requests based on task type, price, and caching potential, increasing the cache hit rate from 5% to an impressive 60%.

This strategic shift has quickly triggered a chain reaction in the Western tech industry. Startup Lindy's CEO has recently switched to using Deepseek v4, while cloud computing giant Snowflake is also intensively testing Chinese models, positioning them as a cost-effective alternative to OpenAI and Anthropic. This move directly puts pressure on Western top AI labs preparing for their IPOs, posing a severe test to the valuation systems that rely on high unit prices to maintain growth expectations.

QQ20260629-085639.jpg

With high-performance reasoning models like GPT-5.x-Thinking and Opus4.5 being launched in the market, enterprises' consumption of tokens is experiencing explosive growth. Under the "token maximization" trend tolerated by giants like Amazon and Meta, where employees are allowed to consume tokens without limits, Coinbase has introduced a more rational accountability mechanism, requiring AI investment to be tied to actual business output.