Thoughts After Seeing Nvidia's Earnings Release
The results themselves apparently beat expectations again. But now the key thing to watch feels like it has shifted from "how much more" to "when AI infrastructure investment will slow down."
With data center revenue now such a large share, even capex plan announcements from a few big tech companies make the stock price swing. I think this has become less about earnings and more like a bet on the cycle.
Of course, I can't judge any of this. I'm just watching and learning.
5 answers
Agreed, it’s a cycle bet. It drops 5% on a single line of announcement wording rather than on earnings—how is that an earnings stock?
But even if it's Big Tech CAPEX plans, in the end it's money that flows right back to Nvidia lol. It's a circular structure, so I'm not sure it's really demand. It basically comes down to: if Microsoft doesn't buy servers, what is OpenAI supposed to run on?
Every time I look at earnings, I find myself only checking the data center revenue share chart. Once that share goes above 90%, I can’t help but wonder whether the other business units become meaningless... I’m just here to look at the numbers, so I can’t really judge, but the post was well organized, so I’m saving it.
But can we really be that sure about an "AI infrastructure slowdown"? Inference demand is only just beginning, and there's also the chip generation replacement cycle—replacement demand alone seems like it could sustain things for years. I can't help thinking this is just dismissing the cycle without evidence.
As the author said, it's true that the point is 'when it breaks' rather than 'how much more,' but the problem is that nobody can get that timing right. In 2024, too, people who were sitting on the sidelines saying that all missed out, didn't they? Ultimately, you can only confirm whether that cycle is really breaking after the stock price has dropped -30%, and before that, there's no answer.