Coinbase is adopting Chinese AI models like GLM 5.2 and Kimi 2.7 to reduce expenses. The shift demonstrates growing pressure on Western AI labs to compete on pricing.
CEO Brian Armstrong announced the company switched to Chinese alternatives after implementing an automated routing system that selects the most cost-effective model for each task.
The move delivered significant savings. Coinbase halved its AI spending while token usage continued climbing, suggesting improved efficiency rather than reduced usage.
A key optimization came from better caching, which boosted the hit rate from 5 to 60 percent. This means the system can reuse previous results more effectively, reducing redundant processing costs.
The decision reflects a broader trend in tech. As Western AI providers like OpenAI and Anthropic maintain premium pricing, companies face mounting pressure to explore alternatives. Chinese models, increasingly competitive in capability, offer lower-cost options.
Coinbase's approach—using an intelligent router rather than committing to a single provider—allows flexibility to switch between vendors based on task requirements and pricing changes. The strategy highlights how the AI market is fragmenting beyond dominant Western players.
Small and medium-sized businesses are thriving despite economic headwinds by leveraging AI tools that enable them to compete with larger enterprises, according to Payoneer CEO John Caplan.
Muddy Waters Research CEO Carson Block is reconsidering plans for a long-short fund in India, citing artificial intelligence risks as a key factor. The firm is returning to the drawing board on the initiative.
Netflix has integrated artificial intelligence into roughly 300 titles so far this year, signaling the company's accelerating adoption of AI tools across its content production pipeline.
China's advancing AI capabilities are strengthening President Xi Jinping's position to influence global AI governance, while simultaneously raising security concerns in both Washington and Beijing.