Moxi closed nearly 1B yuan in a C+ round, its second roughly 1B yuan raise in five months, aimed squarely at sparse inference silicon. DeepSeek-V4-Pro activates only ~49B of its 1.6T params per token,

What happened: Moxin AI (墨芯) completed a C+ funding round of nearly 1 billion yuan, its second round of roughly that size in five months, as it develops sparse inference chips. Key numbers: Nearly 1 billion yuan C+ round; prior round of nearly 1 billion yuan 5 months earlier; founded 2018; DeepSeek-V4-Pro has about 1.6 trillion total Why it matters: The content frames investor interest as recognition of Moxin's sparse inference chip technology and commercialization, and positions sparse computing as a differentiated path for Chinese AI chips as model costs and energy use constrain scaling.

Moxin AI (墨芯) completed a C+ funding round of nearly 1 billion yuan, its second round of roughly that size in five months, as it develops sparse inference chips.

The content frames investor interest as recognition of Moxin's sparse inference chip technology and commercialization, and positions sparse computing as a differentiated path for Chinese AI chips as model costs and energy use constrain scaling.