-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
GPU prices are bound to fall, AI training costs likely to drop
- Thread starter Zarax
- Start date
Nvidia Stock May Fall As DeepSeek’s ‘Amazing’ AI Model Disrupts OpenAI
Barrett Woodside, co-founder of the San Francisco AI hardware company Positron, said he and his colleagues have been abuzz about DeepSeek.www.forbes.com
If true, the AI bubble is going to deflate soon at least for needed GPU horsepower.
In simple terms, if datacenter GPU demand drops the consumer segment will increase its relative importance and will have less competition from higher margin products.
Now let's see if they can pull another trick from the leather jacket sleeves...
Nvidia's share price has started to reverse the losses
Investors are realising it's not as bad as it seems, or at least it makes sense to wait for info.
Deepseek claims their R1 AI model cost $6 million to build, but the problem is they used to be a crypto mining company and already have acccess to 50,000 A100 GPUs.
If I was to go buy 50,000 A100 GPus today, it would cost me approximately $1 billion USD. So this $6 million claim is dubious
Last edited:
You need to get them at Jensen MSRP, not OCUK MSRP.
We willl see, as consumers lower datacenter demand could be beneficial, unless of course the next bubble (digital twins) will raise sooner than expected, but that shouldn't be as impactful.
Demand won’t drop, especially since most consumers are these multi-billion dollar companies.
I have numerous gripes over all the AI BS that gets rammed down our throats, but there is value, especially when more targeted demand is identified.
What might affect Nvidia is when other companies get their own chips and libraries built.
If training costs fall because you suddenly need half or less of the projected GPUs then billion dollar companies WILL lower demand.
You don't grow without being ruthlessly efficient about cutting costs first.
They all switch to buying the cheaper gaming cards rather than the 40k enterprise GPUs.Nvidia Stock May Fall As DeepSeek’s ‘Amazing’ AI Model Disrupts OpenAI
Barrett Woodside, co-founder of the San Francisco AI hardware company Positron, said he and his colleagues have been abuzz about DeepSeek.
If true, the AI bubble is going to deflate soon at least for needed GPU horsepower.
In simple terms, if datacenter GPU demand drops the consumer segment will increase its relative importance and will have less competition from higher margin products.
Now let's see if they can pull another trick from the leather jacket sleeves...
Last edited:
And have you seen how it responds in their app to questions about Chinese president etc.?
It's 100% gov. thing, even in the West adhering to all Chinese legal requirements and their official gov. policies. That said, even if it requires less computational power, it'll just see wider used but not less data centres hardware used - as in, it could actually be profitable as currently the whole AI industry is driving on hype and not profit. Aside NVIDIA everyone else is losing a ton of monies, including OpenAI, because it's just too expensive to run currently.
Last edited:
Yea but there’s no proof that that is the case, and it’s unlikely that some “tiny” Chinese company with almost no info, has miraculously halved hardware requirements.
What’s been discussed is that Deepseek uses a different approach. It’s cheaper to train because they’re doing less training, but it’s more expensive to query as it has a heavier thinking cycle.
Still way too early to draw any real conclusions.
Nvidia has released a statement
DeepSeek are violating first law of Huang. The only way savings can be achieved is by buying more Nvidia GPUs
DeepSeek are violating first law of Huang. The only way savings can be achieved is by buying more Nvidia GPUs
It's funny how Nvidia's share price crashed before anyone has had a chance to actually test and reproduce the "trained on only 6 million USD cost" model by DeepSeek.
Nvidia share prices: Pumped by hype, Dumped by fear
Not sure how genuine the Chinese claim is. Apparently they were running it on un-licenced nvidia hardware and some of the answers it gives are identical to ChatGPT and other seem tailored (for political reasons) lol. So businesses aren't going to trust it.
Nvidia isn't going to stay low, this is only one part of their business. Good time to buy if anything.
Nvidia isn't going to stay low, this is only one part of their business. Good time to buy if anything.
Last edited:
That's brilliant
First law of Huang - more you buy, more you save I wont be surprised if deepseek in a background asking chatgpt and then send answer back to user
That's brilliant
I wont be surprised if deepseek in a background asking chatgpt and then send answer back to user
Yea, though the CCP filter...
That’s what that stupid rabbit thing was doing. Just a cheap Android phone hooked up to ChatGPT.That's brilliant
I wont be surprised if deepseek in a background asking chatgpt and then send answer back to user
GPU prices won't fall because profit margins are too low
NVIDIA's tight GeForce RTX 50 margins put pressure on board partners: 'MSRP feels like charity' - VideoCardz.com
NVIDIA makes it impossible to sell RTX 50 cards at MSRP Where are RTX 50 cards at MSRP? GeForce reviews are typically split between MSRP and non-MSRP models. This was NVIDIA’s way of encouraging board partners to launch at least one model at the suggested retail price, and board partners usually...videocardz.com
Keep in mind those are referring to profit margins for AIB partner companies like MSI and Zotac, not Nvidia themselves.
Nvidia didn't become a trillion dollar company by running low profit margins.
Last edited: