Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Nvidia’s $589B DeepSeek rout

finance.yahoo.com

Post reply on HN
Print view View original

231–240 of 1001 posts

Re: Nvidia’s $589B DeepSeek rout

#231

Post by aurareturn »

Deepseek should cause Nvidia and TSMC stocks to go up, not down. I'm buying more Nvidia and TSMC today.

1. More efficient LLMs should lead to more usage, which means more AI chip demand. Jevon's Paradox.

2. To build a moat, OpenAI and American AI companies need to up their datacenter spending even more.

3. DeepSeek's breakthrough is in distilling models. You still need a ton of compute to train the foundational model to distill.

4. DeepSeek's conclusion in their paper says more compute is needed for next break through.

5. DeepSeek's model is trained on GPT4o/Sonnet outputs. Again, this reaffirms the fact that in order to take the next step, you need to continue to train better models. Better models will generate better data for next-gen models.

I think DeepSeek hurts OpenAI/Anthropic/Google/Microsoft. I think DeepSeek helps TSMC/Nvidia.

Re: Nvidia’s $589B DeepSeek rout

#232

Post by NicoJuicy »

aurareturn wrote: post #231

Deepseek should cause Nvidia and TSMC stocks to go up, not down. I'm buying more Nvidia and TSMC today. 1. More efficient LLMs should lead to more usage, which means more AI chip demand. Jevon's Paradox. 2. To build a moat, OpenAI and American AI companies need to up their datacenter spending even more. 3. DeepSeek's breakthrough is in distilling models. You still need a ton of compute to train the foundational model…

Ugh. No, we can now run a decent model through CPU. Not an expensive video card.

Just try out the standard deepseek-r1 or even the deepseek-r1:1.5B through ollama.

No need for expensive hardware anymore locally. My PC ( without Nvidia card/expensive hardware) runs the a deepseek 1.5 b query fast enough - 2 - 9 seconds until it's finished.

Re: Nvidia’s $589B DeepSeek rout

#233

Post by bufferoverflow »

nabla9 wrote: post #128
esperent wrote: post #13

Earlier quoted context omitted.

> DeepSeek spends $6 million in old H800 hardware to develop open source model that overtakes ChatGPT. DeepSeek claims that's what they spent. They're under a trade embargo, and if they had access to any more than that it would have been obtained illegally. They might be telling the truth, but let's wait until someone else replicates it before we fully accept it.

Huggingface is currently replicating it. Replications of small models indicate that they don't lie any significant amount. The architecture is cheap to train. Berkeley Researchers Replicate DeepSeek R1's Core Tech for Just $30: A Small Model RL Revolution https://xyzlabs.substack.com/p/berkeley-researchers-replicat...

You call R1 a small model? It's a 671-billion parameter model.

Re: Nvidia’s $589B DeepSeek rout

#234

Post by Jlagreen »

blitzar wrote: post #198
jamil7 wrote: post #8

Shouldn't this news be bullish long-term for them?

The $500 billion data center can now be $50 billion. That is excellent news, unless you were the company that was expected to sell $450 billion of GPUs with a 95% gross margin to that project.

$50 billion can be afforded by WAY WAY more sites and companies so Nvidia will simply delivered to >10x more data centers. Or instead of shipping 100k GPUs to Meta, they will ship 100k GPUs to 10 different customers.

For Nvidia, it is great news because now finally, the concentration of GPUs at Hyperscalers will end and every Fortune company can finally get their local data center to train their AI Models.

Because if training AI models becomes more efficient and easier then the ones being that business are at risk so basically Big Tech. Nvidia isn't in that business but in the business of providing tools to train.

Fortunately, Big Tech can easily do something to prevent ANYONE for competing. They simply buy all available GPUs. Oh wait, haven't they been doing it for years? Excatly!

People really don't get what an arms race and market competition race is.

How do I prevent disruption? I simply buy all the tools the competition needs to disrupt me.

See, if Fortune 500 companies want to build large data centers but can't because all Hyperscalers buy the GPUs then eventually they will rent from cloud as otherwise they can't get the GPUs.

Re: Nvidia’s $589B DeepSeek rout

#235

Post by aoeusnth1 »

A huge leap in the cost effectiveness of AI capabilities only paves the way for faster timelines to ASI. I'm not sure why that would reduce the economic value of Nvidia. Pretty sure this is a reaction in the wrong direction. Nvidia should be popping.

Re: Nvidia’s $589B DeepSeek rout

#236

Post by aurareturn »

NicoJuicy wrote: post #232
aurareturn wrote: post #231

Deepseek should cause Nvidia and TSMC stocks to go up, not down. I'm buying more Nvidia and TSMC today. 1. More efficient LLMs should lead to more usage, which means more AI chip demand. Jevon's Paradox. 2. To build a moat, OpenAI and American AI companies need to up their datacenter spending even more. 3. DeepSeek's breakthrough is in distilling models. You still need a ton of compute to train the foundational model…

Ugh. No, we can now run a decent model through CPU. Not an expensive video card. Just try out the standard deepseek-r1 or even the deepseek-r1:1.5B through ollama. No need for expensive hardware anymore locally. My PC ( without Nvidia card/expensive hardware) runs the a deepseek 1.5 b query fast enough - 2 - 9 seconds until it's finished.

The world isn't satisfied with a "decent model". The world is trying to reach AGI.

Further more, reasoning models require more tokens. The faster the GPU, the more thinking it can do in a set amount of time. This means the faster the hardware, the smarter the model output. Again, that reinforces the need for faster hardware.

More thinking = smarter models

Faster hardware = more thinking

Therefore, faster hardware = smarter models

Re: Nvidia’s $589B DeepSeek rout

#237

Post by Jlagreen »

JSR_FDED wrote: post #63

I saw the interview with Dario Amodei. Look at 0:48 https://m.youtube.com/watch?v=7LNyUbii0zw He seems adamant that there are no diminishing returns to scaling AI. I don’t want to stir up conspiracy theories but I do think that currently all the big AI players have a vested interest in the message that the current scaling paradigm is the right one, and that this is a supremacy issue wrt China. It drives so much inves…

The issue is it feels like we came to a stop but Hyperscalers are simply waiting for Blackwell. That's all.

Why buy 100k Hoppers if 20k Blackwell offer the same compute so then it's better to buy 100k Blackwells right?

Backwell will increase cluster scaling easily by 10x performance and if you buy 10x of them then your compute on a cluster will be 100x than before. If it takes you to wait 6-12 months for that then so be it. You will easily make up the time in the end with the speed up.

Re: Nvidia’s $589B DeepSeek rout

#238

Post by surgical_fire »

esperent wrote: post #207
maxglute wrote: post #167

Earlier quoted context omitted.

>been obtained illegally. PRC companies breaking US export control laws is legal (for PRC companies). Maybe they're trying to avoid US entity listing, lot's of PRC companies keep mum about growing capabilites to do so. But the mere fact Deepseek is publicizing means they're unlikely to care about the political heat that is coming and the ramifications. If anything, getting on US entity list probably locks in their em…

> PRC companies breaking US export control laws is legal So long as they don't plan to do any business with the US or any of their allies I guess.

Which allies? The ones the current US president is threatening in all sorts of manner?

I actually hope he doubles down. I would love for EU to rely less on the US. It would also reduce the reach of the silly embargoes that benefit no one but the US.

Re: Nvidia’s $589B DeepSeek rout

#239

Post by mullingitover »

aurareturn wrote: post #231

Deepseek should cause Nvidia and TSMC stocks to go up, not down. I'm buying more Nvidia and TSMC today. 1. More efficient LLMs should lead to more usage, which means more AI chip demand. Jevon's Paradox. 2. To build a moat, OpenAI and American AI companies need to up their datacenter spending even more. 3. DeepSeek's breakthrough is in distilling models. You still need a ton of compute to train the foundational model…

That’s a rational stance. However, the Buffet Indicator is flashing red, we’re in a dotcom-era sized bubble, and it only takes a little bit of ill-founded worry to kick off a serious panic.

Re: Nvidia’s $589B DeepSeek rout

#240

Post by refurb »

rhubarbtree wrote: post #214
refurb wrote: post #108

Earlier quoted context omitted.

> Your agenda collapses and you enter reactive mode. Where has Trump's agenda collapsed? I might have missed the press release. And why would a curveball on AI throw off an agenda around trade, immigration and military engagement? I don't follow. China could take the lead on AI and I don't see it would impact any of those things. Isn't DeepSeek open source? The US already has access to it, so what leverage could Chin…

> Where has Trump's agenda collapsed? I might have missed the press release. Well, it's too early to say if it has happened in this case. But we've seen it happen again and again, so it will not be a surprise if it happens. > And why would a curveball on AI throw off an agenda around trade, immigration and military engagement? I don't follow. Well, trade restrictions against China have just backfired spectacularly. S…

> Well, it's too early to say if it has happened in this case

Ah ok, so you’re just guessing. You wrote it as if it had already happened.

And I think you’re confusing the stock market with the broader economy. The economy as a whole is unaffected by AI at this point.

And the trade war with China involved hundreds of industries beyond AI. Most of it is manufacturing. I’m not sure how failure of AI sanctions (questionable conclusion) somehow means the trade issues around machine parts needs to be abandoned.

I agree China has motive but I’ve heard so many claims that the Chinese government doesn’t control research or businesses like TikTok.

Post reply on HN