Vals analysis: open-weight models performing multi-stage tasks, like building a web app, can have an environmental impact 10K times greater than simple queries (Bloomberg)
Benchmarker Vals AI finds that multi-stage tasks can require models to use exponentially more energy or water than simple queries do.
Chinese AI companies are rapidly closing the gap with their US rivals. According to analysts, the global artificial intelligence boom has reached a point where autonomous agents now consume more data than human users, leading to a surge in processing costs. This shift has altered the economics of AI deployment, giving lower-priced Chinese models a competitive edge.
The increased consumption is driven by autonomous agents that operate independently to execute multi-step workflows, such as writing software or managing business operations. These agents trigger a cascade of automated model calls, resulting in significantly higher token usage. Data from OpenRouter showed that daily token consumption from agentic workloads hit 7.3 trillion in early August, 14 times the level six months earlier.
This development has implications for the competitive landscape of AI, with Chinese AI companies potentially benefiting from the changed economics. The US has imposed restrictions on China's access to advanced chips, but Chinese AI companies appear to be adapting to these limitations.
Brief written by urgent.news from Techmeme, SCMP Tech — 2 reports on this story. Machine-written — may contain errors; check the original before relying on it.
