Kevin Kelly says low-cost tokens and open-source models could aid China AI

Speaking at the 2026 World Artificial Intelligence Conference on July 18, the futurist said cutting model costs could be disruptive, though open-source development still needs substantial funding.

Summary

Token pricing is set to matter more in artificial intelligence, and lower-cost open-source models may give China’s AI sector an edge, futurist Kevin Kelly said at the 2026 World Artificial Intelligence Conference on July 18. He said a model priced at one-tenth of Anthropic’s could prove disruptive, underscoring how cost efficiency can shape competition as companies seek wider adoption. Kelly also cautioned that open-source development does not remove the need for heavy investment, because large models remain expensive to build.

Terms & Concepts
  • open-source models: AI models with publicly accessible code
  • tokens: Units used to measure AI text processing
  • large models: AI systems trained with vast data