BlockBeats News, July 21,

Amidst the recent significant volatility in the US stock market's AI Chain, the market briefly dubbed Kimi K3 as the new "DeepSeek moment": a low-cost, high-performance, openly-weighted Chinese model that reignited investors' concerns about the ROI of US AI capital expenditure, model pricing power, and computational demand. The WSJ reported that Moonshot AI's new model has exacerbated chip investors' anxiety.

In its latest report on China's artificial intelligence industry, JPMorgan Chase indicated that Kimi K3 has indeed altered the valuation framework of leading-edge model companies in China. The report stated that after the release of K3, the market began to doubt the long-term competitiveness of companies such as ZhiPu and MiniMax, hence reducing their valuation multiples from a previous approximately 30 times the 2030 expected P/ARR to 20 times.

However, JPMorgan Chase believes that the over 50% correction in ZhiPu's stock price has already excessively reflected this pressure. The firm maintained a "overweight" rating on ZhiPu, while revising the target price from HK$2400 to HK$1600; MiniMax retained a "neutral" rating, with the target price lowered from HK$240 to HK$160.

The core assessment of the report is that K3 has narrowed the competitive window of GLM-5.2 but has not excluded ZhiPu from the ranks of China's leading-edge models. JPMorgan Chase stated that GLM-5.2 still remains among the top production models in China, and ZhiPu will reassert its competitiveness in the coming months through GLM-5.3 and the 2T+ flagship model. If the new models can re-enter the leading edge in coding, inference, and intelligent agent tasks, ZhiPu still has the capability to sustain its business growth.

More importantly, the commercialization of Chinese models is still in its early stages. JPMorgan Chase estimates that ZhiPu's latest indicative ARR is around $1 billion, DeepSeek is around $500 million, MiniMax and Kimi are approximately $300 million each, and the total for leading independent model suppliers in China is around $2.1 billion; in comparison, Anthropic's ARR has already reached around $69 billion. This indicates that the revenue scale of Chinese model companies is still significantly undersized, and the potential for future growth has not been completely squeezed out by the emergence of Kimi K3.

Kimi K3 has also left another signal to the market: stronger domestic models are attempting to prove their capabilities at a higher price. The report mentioned that the pricing of K3 API is significantly higher than the previous generation K2.7 Code, reflecting its enhanced capability in high-value tasks such as coding. If customers are willing to pay for stronger inference, longer context, and programming capabilities, the business model of Chinese model companies will shift from price competition to capability-based pricing.

Therefore, the tone of this report is not pessimistic. The Kimi K3 brought a valuation reassessment and magnified concerns about the US stock AI chain's "Cheap China Model"; however, in J.P. Morgan's view, this impact also demonstrates that the Chinese model layer is approaching the global frontier. For Wisetune, the next round of model release will be a key repricing window.