TechFlow news, According to Chaoxiang Research, Goldman Sachs pointed out in a July 18 report that Moonshot AI released the Kimi K3 model, with 2.8 trillion parameters, topping the Arena.ai coding leaderboard by surpassing Claude Fable 5 and GPT-5.6 Sol, with API pricing at $2.3 per million tokens, setting a new high for Chinese models. Two days after the release, Zhipu AI fell 28%, MiniMax fell 16%, Nasdaq 100 index futures fell more than 1.8%, and the Philadelphia Semiconductor Index cumulatively fell more than 18% from its high.
Goldman Sachs believes Kimi K3 marks a turning point: a Chinese laboratory unable to match the largest scale of pre-training computing power in the West, through architectural innovation and reinforcement learning, rapidly narrowed the gap with top US models, proving that "scaling" is no longer the only winning path. Goldman Sachs warned that the "era of computing power expansion" may be ending, and the AI infrastructure investment logic built around "the more computing power, the better" needs to be rewritten. Goldman Sachs maintains a Buy rating for MiniMax and Neutral for Zhipu AI. Subsequently, watch for the intensive launch of 2-5 trillion parameter models such as Zhipu GLM, Alibaba Qwen, and MiniMax M3 Pro.
