Moonshot AI kicks off pre-IPO funding round at $50 billion target valuation
Moonshot AI began pre-IPO funding negotiations on August 15 at a target valuation of $50 billion (~RMB 340 billion), racing DeepSeek's reopened Series B for capital.
Times Weekly reported on August 15 that Moonshot AI has started pre-IPO final-round funding negotiations at a target pre-money valuation of $50 billion (~RMB 340 billion). The round races DeepSeek's reopened Series B for Chinese large-model capital.
Moonshot's flagship Kimi K3 was open-sourced on July 16 with 2.8 trillion parameters (104 billion active), priced at $3/$15 per million input/output tokens, with a 1M context window. It is the largest open-source MoE model and leads Claude Opus 4.8 (56 points) on the Artificial Analysis Intelligence Index at 57 points.
Kimi K3 is now running on Alibaba Cloud's Lingjun Zhenwu M890 supernode instance, the first domestic supernode running 2-trillion-parameter-plus models. Moonshot founder Yang Zhilin rejected offers from Apple and overseas academic positions to return to China to found the company. Kimi K3 is the key milestone of his team's open-source strategy. Kimi K3 is currently being benchmarked alongside Qwen3.8-Max, DeepSeek V4 Pro, Meta Muse Glimmer, and NVIDIA Nemotron 3.5.