Breakthroughs in cheap Chinese open-weight artificial intelligence models have spooked US investors, but analysts argue plummeting model costs will benefit the AI industry in the long run by supercharging global demand for AI systems.Companies across the AI industry have slashed prices in recent weeks, with large-language model (LLM) inference prices per million tokens falling from above US$2 at the start of June to just US$1.2 this week, according to research firm Silicon Data’s LLM Token Expenditure Index, which tracks both frontier providers and open-weight platforms.Under pressure from cheaper Chinese models, Silicon Valley firms have cut the prices of closed models in an effort to grab market share. OpenAI, for instance, last week announced an 80 per cent discount on developer pricing for its lightweight GPT-5.6 Luna model, and a 20 per cent discount on the mid-tier GPT-5.6 Terra.That contributed to a severe AI stock sell-off last month, amid investor concerns that US hyperscalers were overvalued. But analysts say the intense competition will benefit the AI industry in the long run.The booth of Chinese AI model company MiniMax seen at the World AI Conference in Shanghai on July 17, 2026. Photo: AP“Competition is up and prices are down,” Silicon Data wrote on social media platform X on Wednesday. “This is good for consumer and enterprise users of AI (agents) and promotes much wider and faster AI adoption.”
Jevons Paradox: why China’s cheap AI models could be good for Silicon Valley
China’s open-weight models have led AI firms to slash prices, but they will also massively boost overall AI demand, analysts say.
Chinese models reduced LLM inference from $2+ to $1.2/M tokens, forcing OpenAI to cut pricing 80% on GPT-5.6 Luna. Lower costs accelerate AI adoption (Jevons Paradox), driving enterprise AI and agent deployment growth despite margin compression.









