A provocative new theory from Decagon CEO Jesse Zhang suggests that the relationship between frontier AI models and open source alternatives is not one of direct competition, but rather two phases of a single life cycle. Zhang argues that expensive frontier models are used to prove out new use cases, which are then passed along to cheaper open source alternatives as they mature. While mature deployments switch to lighter models, new use cases keep arising, meaning overall spend on frontier models barely goes down. Data from platforms like Vercel and OpenRouter supports this idea. While DeepSeek has surged in token volume, Anthropic still accounts for more than half of overall AI spend on Vercel, and Opus 4.8, despite processing fewer tokens than DeepSeek V4 Flash, likely captures the lion's share of spending due to its much higher cost per token. The figures suggest frontier labs like Anthropic aren't suffering from the rise of open source, as the market for AI-addressable tasks is growing so fast that top models maintain their position by dominating early-stage deployments.
AI-Powered News Curation