Their stock price will tank. They have a $3B market cap because they’re selling shovels in a gold rush. Once the gold rush is over, that valuation will go back to where they were three years ago. Probably lower, because the stock market tends to overcorrect on these things.
Companies base their capital on their stock price, and a drop like that can kill companies. Doesn’t mean for sure that Nvidia will die, but they could.
Like he said though, being fabless means there’s no assets to hold the bag on. It’s the fabs that stop getting profitable orders that tank. Worst case is they do layoffs.
Their fundamentals are too strong. They have market dominance with extremely steady technological progress against really bad competition. LLMs aren’t going to disappear when the shitty overpromising bubble pops. Generative AI isn’t going anywhere. Any of the thousands of other uses for their raw power are still there. They’ll just be at the ground floor of whatever the next math heavy hype cycle is, just like they were with crypto and LLMs, because cuda is the best way to get shit done, whether what you’re doing is useful or trash.
Their stock price will tank. They have a $3B market cap because they’re selling shovels in a gold rush. Once the gold rush is over, that valuation will go back to where they were three years ago. Probably lower, because the stock market tends to overcorrect on these things.
Companies base their capital on their stock price, and a drop like that can kill companies. Doesn’t mean for sure that Nvidia will die, but they could.
Like he said though, being fabless means there’s no assets to hold the bag on. It’s the fabs that stop getting profitable orders that tank. Worst case is they do layoffs.
Their fundamentals are too strong. They have market dominance with extremely steady technological progress against really bad competition. LLMs aren’t going to disappear when the shitty overpromising bubble pops. Generative AI isn’t going anywhere. Any of the thousands of other uses for their raw power are still there. They’ll just be at the ground floor of whatever the next math heavy hype cycle is, just like they were with crypto and LLMs, because cuda is the best way to get shit done, whether what you’re doing is useful or trash.