Chinese AI Model Kimi K3 Suspends New Subscriptions Due to High Demand
HONG KONG — The newly launched Chinese AI model, Kimi K3, has taken the U.S. tech world by surprise, but it has now paused new subscriptions after being overwhelmed by demand shortly after its release.
This situation underscores the challenges faced by Chinese AI companies in meeting the needs of a growing user base both domestically and internationally, especially as the tech competition between the U.S. and China intensifies.
Advanced Chinese open-source AI models, like DeepSeek’s latest V4, are rapidly gaining recognition worldwide. In early 2025, DeepSeek made waves in the global market, establishing China as a serious player in AI, particularly since many Chinese frontier models are open source and available for public use.
“Kimi K3 has attracted much more attention than we anticipated,” stated Moonshot AI, a Beijing-based company, in a post on X. “In just the last 48 hours, demand has nearly reached our current capacity.”
Moonshot indicated that it is focusing on current subscribers and will pause new ones temporarily. “We’re working to increase capacity quickly and will reopen subscriptions in batches,” they added. A similar notice was shared on Chinese social media platforms.
The public launch of K3, which seeks to compete against advanced models like those from Anthropic and OpenAI, has unsettled competitors in the U.S. With 2.8 trillion parameters, it is considered the largest open-source AI model in existence.
Lian Jye Su, a chief analyst at technology research firm Omdia, noted that new AI releases usually generate significant interest, which can put a strain on existing computing resources. “This indicates that Moonshot AI may lack enough compute chips to meet the current demand surge,” he said.
Su suggested that the main reason for this situation is likely Moonshot’s unexpected underestimation of K3’s appeal. K3’s “very demanding” computing needs make resource allocation tricky and costly, he added.
In terms of coding capability, a performance measurement for AI models, Kimi K3 has emerged as a leader among major advanced models following its release on the evaluation platform Arena.
The introduction of K3 has also impacted the stock prices of major American tech companies, as concerns grow about how affordable Chinese models might affect pricing and demand in the U.S. market, even with existing tech restrictions limiting China’s access to top-tier technology, including cutting-edge chips.
K3 is part of a wave of newer Chinese AI models that have received significant global attention.
Over the weekend, Alibaba unveiled its own advanced AI model, Qwen3.8 Max, boasting 2.4 trillion parameters, making it one of the most powerful, ranked just behind Anthropic’s Fable 5. Additionally, in the previous month, the startup Zhipu, also known as Z.ai, introduced its GLM-5.2 model, which has been rapidly adopted worldwide.
