US-China AI Model Cycles Shrink to 44 Days as Autonomous R&D Accelerates

New AI models are now being released in the US and China every 44 days on average, roughly one-third of the previous development cycle, according to a Nikkei report summarized by Dick Weisinger on September 21, 2026. The acceleration is largely driven by AI systems taking on significant roles in R&D, producing a sharp rise in code generation and performance gains, which is intensifying concerns that self-evolving AI could become uncontrollable. The shortened cycle carries implications for national security, economic competitiveness, and the global regulatory landscape amid the US-China technological rivalry.

US-China AI Model Cycles Shrink to 44 Days as Autonomous R&D Accelerates AI development in the US and China is accelerating at an unprecedented pace, with new models now being released every 44 days on average, a third of the previous development time. Dick Weisinger · September 21, 2026 · 2 min read · Source: nikkei jp · Issue 103 East Asian Technology Intelligence Japan & China tech news — translated, contextualized, and delivered for Western readers. Free. Unsubscribe anytime. This story ran in Issue 103, alongside three other stories. AI & Machine Learning AI development in the US and China is accelerating at an unprecedented pace, with new models now being released every 44 days on average, a third of the previous development time. This rapid acceleration is largely driven by AI systems themselves taking on significant roles in R&D, leading to a dramatic increase in code generation and performance improvements. The self-evolution of AI is intensifying concerns about its potential to become uncontrollable. The speed of AI model development has significant implications for national security, economic competitiveness, and the global regulatory landscape, particularly in the ongoing technological rivalry between the US and China. The debate around AI ‘threat theory’ or ‘x-risk’ is prominent in Western discourse, focusing on the existential dangers of advanced AI.