On September 20, multiple outlets reported that Zhipu has closed approximately $5 billion in financing—including about $2 billion in share placement and about $3 billion in convertible bonds—a funding scale described as a record for China's AI industry. The proceeds will go mainly to next-generation GLM models, a fully self-trained system and computing infrastructure.
The same day, Zhipu launched its new lightweight model GLM-5.3-FlashX, which peaks at 200 tokens per second—about 5x faster than GLM-5.3-Flash—while keeping pricing in the same tier. It carries 320B total parameters with only 18B activated, uses a sparse-plus-linear-attention hybrid architecture, and runs entirely on a domestic AI accelerator cluster of over 100,000 cards.
The financing underscores investor conviction that China's foundation-model players are moving from model-building into defensible, large-scale deployment, as AI applications accelerate from technology validation toward scenario penetration and commercial closed loops.
Source: China Finance Online / National Business Daily
