What is Moonshot?

Nathaniel Whittemore on inference capacity as the binding constraint for cheap AI models.

“It doesn't really matter if your model is cheap if you don't have the inference to serve it. Moonshot was completely tapped out of compute on opening weekend.”

Moonshot, the Chinese AI company, sold out its inference capacity on opening weekend, signaling strong demand but supply constraints that undercut the appeal of cheap model pricing.

“It doesn't really matter if your model is cheap if you don't have the inference to serve it. Moonshot was completely tapped out of compute on opening weekend.”

US open-source models from Nvidia (Nemotron) and Google (Gemma) are not sufficient substitutes for Chinese open-weight models like Deepseek, Qwen, and Moonshot, leaving startups avoiding paid APIs with no reliable fallback.

“The models from Nvidia, the Neatron family, that the models from Google's Gemma family aren't sufficient to replace what we have from Deepseek, from Zippu, from Moonshot, from Quinn, and we're going to end up in a place where a lot of people are betting on rolling their own, especially startups that don't want to pay open margin for them, and they're just not going to have something to fall back on.”

Moonshot AI raised $3.5B at a $35B valuation and could become a default enterprise model if underutilized frontier compute is redirected to serve it at steep token discounts.

“It's extremely possible that perhaps this wonderful company called Moonshot, which we talked about 3 seconds ago, could be the model of choice and that somebody's going to take that compute which is not going to be used by Frontier LM and put Moonshot on it and sell it to enterprises at 10 cents a dollar on the 10 cents a dollar tokens.”