source: techcrunch ai: the real ai race may no longer be at the frontier
level: business
chinese open-weight models accounted for 41% of downloads on hugging face this spring, surpassing u.s. models. on openrouter, the top six most popular models are all open models from chinese firms like tencent, xiaomi, and deepseek. anthropic's claude opus 4.7 trails in seventh place. data from vercel shows open-weight models handled nearly a third of ai requests in june, absorbing much of the volume-heavy infrastructure while closed models serve as a higher-cost premium layer.
hugging face ceo clem delangue says companies increasingly prefer owning their ai models rather than renting them, especially after seeing the costs of scaling closed frontier models. a new repository is created every seven seconds on the platform, which hosts almost three million public models. half of fortune 500 firms use hugging face to deploy private and open source models, reflecting a trend toward using many customized models instead of a single dominant one.
the rise of open models has sparked debate over safety and control. anthropic's dario amodei warns that scaling powerful open model weights could be dangerous, while delangue argues that concentration of power is the biggest risk. he says transparency from open models helps defenders patch cybersecurity risks and that keeping models closed creates asymmetry of power. microsoft ceo satya nadella also warned against single provider lock-in, urging firms to control their own learning loops.
why it matters: the shift to open models changes how ai is deployed, making it cheaper and more customizable, but raises questions about safety and market power.
source: techcrunch ai: the real ai race may no longer be at the frontier