1mo ago

Why did Nvidia stock drop with R1 hype?

I get the stupid basic excuse about big tech purchasing loads of silicon. That reasoning seems deeply flawed and idiotic to me. If Deepseek R1 democratizes training more for less, then that means customers with mid to small size data centers like universities now have access to train and research models in this space. Nvidia does not have a real competitor, so they get the sale. Their potential customer base just grew exponentially right? OpenAI should be devalued massively by this change, but I don't see why anything impacts Nvidia negatively in this instance. Am I missing something or is the market this level of stupid? (I have no skin in this game)

33 comments

In a gold rush, be the guy selling shovels not looking for gold.
OpenAI: we made a LLM look how good it is.
Investors: wow, here is money for more, how did you do it?
NVIDIA: we sold them the hardware to do it. Our hardware is the best, expensive to buy, and had large markups.
Investors: awww yea, time to make money, buy buy buy.
Chinese AI: we made an AI without NVIDIA hardware*.
NVIDIA/investors: wait what?
Investors: sell sell sell
- They used Nvidia hardware though. Just less of it.
  
  DeepSeek showed you can do the same work with less cards so Nvidia is going to sell less cards.
  
  They didn't use cuda though and there are reports of AMD chips doing very well on R1.

Because stock traders have no interest in any company or any product. They are all about staying one step ahead of the next trader. As opposed to investors that also have no interest in any company or any product. They are all about the next quarterly report.
This economic system is bullshit.
- It's insane to me how stock prices are basically entirely disconnected from how a company is performing and are dictated by stock market buying and selling pressures.
  You could pick literally any publicly traded company and make its stock price soar just by convincing enough people to buy it, with no relation whatsoever to how the company is performing or forecasted to perform. See: GameStop.
  Nvidia tanked because a bunch of people sold Nvidia stock. Full stop. They may have been motivated by news of deepseek or whatever, but that's not what moved the stock price. Had no one sold it would've stayed exactly where it was.
  Frankly baffling that anyone can look at it and think "yes, this is how it should work and I don't see any problems with it."

The stock market is crazy and it doesn't always act rational. Now that I said this, I think the reason is more a question of Chinese vs US technical superiority. In theory, China has no access to the high technology AI chips from NVIDIA, due to sanctions. This must mean that Deepseek was developed with cheaper chips, and NVIDIA chips are not necessary for training LLMs.
- That is what my intuition was looking for but not finding. I wasn't thinking about the ban and implications. This makes much more sense.
- Sanctions are failing in a huge way to stop chips like the m800 from making it into China. Not only that but the M100 that we are selling to China is essentially a detuned M800. It's hard for investors to keep burning mountains of cash on US AI companies that charge an arm and a leg for their services when a Chinese model comes out totally free and doing just as well as the premium offerings. If investors aren't willing to burn cash, the cash burning dumpster called Nvidia takes a hit. That's why the stock is dropping.

stocks are a confidence game. china showed you can do a lot more with less and i think there was a pretty widely acknowledged awareness that there was already a bubble.. now i'm not a fancy economist but this is what i've gathered
- It’s not a lot more, overall still behind GPT, but it is done with less. So there’s a slight efficiency gain at least, but most of the buzz of R1 is hypeaganda.
  
  This isn't true. R1 trades blows with O1 which is the best model that OpenAI released so far - all the hype on O3 is just vaporware until we have an actual product. Most of the buzz isn't because China made a comparable model with less, it's because they released the weights for free.
  Why would you pay $200/month to OpenAI when you can use R1 for free? Better yet, companies can now self-host it for better security and way cheaper costs. The hype is warranted.

If you were supplying huge amounts and often specified products on contract, you can make a lot of money. DeepSeek just made a lot of the quantity and specifics much less and much more generalised.
I have a good amount of confidence they will do what they did in the crypto boom and pivot to small deployment as the possibility of local LLM's becomes more popular, people will be buying and making dedicated local machines. Now they also have more competitors in this space as the chip embargo to China has caused them to roll their own and so in regards to AI processing Nividia is no longer the only kid on the block.

R1 made waves for two reasons:
It was cheap to train (maybe because they stole/"distilled" openAi's training data).
It is much more efficient to run (it requires much less processing power to run at a decent speed).
I think number 2 is what hit the nVidia stock so hard. Up until the R1 release, the future of AI development would require more GPU and more power. The R1 release showed maybe that isn't the case.
Imagine everyone is out driving around in gas powered cars that get 4MPG. Then a new cheaper gas powered car comes out that gets 60MPG. What happens to the price of gas when the 60MPG car is released? Maybe it fluctuates? Maybe it goes down (less gas is needed due to the more efficient cars)? Maybe it goes up as there are more cars on the road (due to the cheaper entry cost)? What happens if the cheaper more efficient car is electric?
Lots of people see the future of AI as being tied to the hardware and have pumped up nVidia's stock because of that. Anything that goes against that theory will likely let some air out of nVidia's bubble.
- I also think the fact that it's open source kind of takes the mystery out of the ai game and since it's said to be preforming as well as ChatGPT. Takes the wind out of any hardware partners of Openai.

I don't think the stock drop was about R1 hype. I think it was about parent company circumventing NVIDIA CUDA and building AI software in GPU machine code directly. Licensing GPU drivers for cloud services is a major income source that has been threatened
- I think Deepseek was a factor, but so too were the tariffs announced that week and the reviews coming out about their latest line of consumer GPUs, coupled with the fact that they were just so incredibly overvalued to begin with.
  A lot of algorithms likely concluded "It can't go any higher from here" and started selling off.

Market manipulation. GPUs are still heavily required perhaps even more now that u can run a decent mode l localy
- Maybe for training but to just use the model locally for a solo user, a GPU isn't needed at all.
  
  If u wanna run a decently large model u do. The new nvidia card are especially good cos they have much larger amounts of vram.

Fear, uncertainty, doubt. I’ve been buying NVDA through this panic.
- Buy the dip. I more than doubled my position in YNVD on Monday when it was plummeting. As far as I’m concerned, I got a 20% discount on something I was already buying.

This kind of "active" stock trading is essentially an attempt to predict the future. If you own Nvidia stock and hear about a new disruptive technology entering the market, you might predict that its stock value will drop. So, you sell now before the price falls too much, intending to buy back at a discount once you think it has stabilized.
When enough people do this, it triggers a chain reaction - investors see their holdings losing value and rush to sell before it's "too late." But in reality, it’s rarely too late because, more often than not, the price eventually rebounds. The only reason people lose money is that they panic and sell at a loss. There hasn’t been a single stock market crash in history that we haven’t recovered from.
That’s why writing news articles about short-term stock fluctuations is mostly pointless - these movements barely even register on the graph when viewed over the long term.

Disclaimer: not a professional investor, trader, or finance researcher. I literally just have a mix of US and international index funds and bonds.
These are just my personal opinions and I probably don't know more than OP... My view of the stock market, especially with short-term changes is... I don't assume anything to make rational sense, or that the market could really be quite stupid.
Maybe my view was influenced by A Random Walk Down Wall Street... but my main consideration was below:
So-called "meme stocks" and hype cycles have existed for hundreds of years, much longer before the current LLM hype and the previous crypto hype
There is a somewhat "successful" strategy of buying a stock to sell it to the next sucker, which would be very prone to the entire thing collapsing
Especially considering that LLM is the current hype (and the hype cycle might be ending), I... imagine just about anything could happen.
More specifically regarding Nvidia though, I mean they've already made their fair share of money because their technology was used for both crypto and LLMs so... They already won anyway, so who's really hurt from their stock dropping a bit?

Pretty much like the other comment. It was thought that we still needed a lot of AI processing power to advance AI before it became more efficient. Now it turns out R1 does what o1 does but with way less. We'll still need nvidia so I think the stock will rise again. Besides, these are just large language models and are still a ways away from actual AI. I think there will be a second bubble at some point.

The market is not rational. It’s mostly based on vibes and feelings. React first, think later.

I disagree with other commentors that the market is irrational. It appears irrational in relation to the news because, contrary to what we are told, news doesn't move markets, money moves markets (though, yes, some salient times news is what causes the money to move). Many times big news does nothing, and equally many times major movements occur with no news to speak of.
The cause for big money movements into Nvidia seem pretty clear: this past year Nvida has become well known as an foundational buisness for the fastest growing new sector of artificial intelligence. As such, like amazon before it, it has become big enough to swallow it's competitors, which makes it a good long position for investment. So, with good foundations and the lions share of the AI market, when the big investments drive up the price it is easy to get the financial news to drum up retail interest and explain why it is a good place for the average person to put their money, all of which drives the price up further.
It seems to me more likely that this was a semi- (maybe only algorithmically) coordinated dump: it does not make a lot of sense to me that deepseek's model would cause doubt in Nvidia's valuation, as efficient models which briefly beat the leaders seem to pop up from time to time without the same armaggedon and retail investors are generally not keeping up with the bleeding edge of AI developments. Indeed, before the massive sell-off, AI news youtube videos on deepseek were not even particularly well watched.
We are taught that the market reacts to news but, especially with dumps, its seems to me that more often news stories are cooked up as a cover for market movement (see, for instance, the way companies will send news briefs to cnbc to have them help pump and dump stocks).
Maybe it was deekseek, or maybe it was Trump planning to tariff TSMC in Taiwan, or simply that the algorithms decided it was time for a little rug pull on retail.
Whether the whales planed to sell together or the algorithms did I guess is not so important, though it is interesting that CNBC's reporting mentioned that only a week ago big tech investors were discussing deepseek at davos.
Market movements are only out of the ordinary if the big investors lose out. Nvidia, I think (and hope -- because I'm invested), will remain a good long monopolistic stock like amazon, which will continue to have its cycle of good news pumps and bad news dumps as the big players prey on the fomo retail traders.
- I expect Nvidia to boom for at least another 5 years.

I think it's not really rooted in facts. AI is an unsubstantiated hype and the stock market is a bubble. People seemed to have been under the impression, that OpenAI was going to invest several trillions(!) of dollars into Nvidia chips. To me, that always seemed a bit unrealistic. But that's what inflated the Nvidia stock. And now it turns out, to everyones' surprise, that OpenAI isn't the only company who can do AI. And that AI is making advancements and is getting better and more efficient all the time... So that trillion dollar bubble collapses.
To me, that's just silly. AI making progress was the very reason for those people to invest in it. Plus it's not like there is another company manufacturing the chips... Deepseek used Nvidia chips. So IMO they proved they're even better than people previously thought and there is room for improvement... But seems to me the stock market is set on doing it one specific and ineffective way, so it theoretically would need more hardware to do AI.
I think it'll turn out the opposite. The better AI gets, the more it'll get adopted. And that'll lead to more sales, not less. And if Nvidia hardware turned out to be better than we thought, it just proves they're ahead of their competition. So even more reason to invest in them. But the stock market sometimes just does silly things and isn't focused on long term goals.

33 comments