The "battle of the gods" in the era of large models has surprisingly shown some sense of camaraderie. On February 10th, in response to the frequent issues of insufficient computing power recently, Moonshot AI's AI assistant Kimi gave a highly distinctive response.

Overloaded during peak hours, Kimi admitted its lack of computing power.

Recently, many netizens found that Kimi had "intermittent strikes." The page frequently popped up with a message: "Insufficient computing power during peak hours. Switched to K2.5 fast mode. Upgrade your membership for full access to the thinking model." Obviously, even after rapid iteration, Kimi still couldn't withstand the server pressure caused by massive user requests.

A textbook-level public relations move: not only admitting fault but also offering guidance.

Facing user feedback, the official response from Moonshot AI Kimi was both down-to-earth and humorous: "Well, we're looking for computing power. How about using DeepSeek first?" This action quickly sparked heated discussions, with netizens jokingly calling it a "model of mutual support in the AI world."

The new normal in the AI field: too many great products, and computing power is struggling to catch up.

Kimi actively "referring" users to its competitor DeepSeek subtly confirms two things:

  • Surge in user numbers: Domestic large models are experiencing an unprecedented period of user growth.

  • Shortage of computing power: The demand for high-performance computing resources remains a common challenge for all players in the large model arena.