arkhamguy123
Member
Serious question: Should I invest?
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Nvidia hits 1 trillion market value
- Thread starter Buggy Loop
- Start date
Silver Wattle
Gold Member
The AI bubble in effect.
The revenue they currently have and are projected to have in the mid term say they are massively overvalued (PPE), NFI why investors think their revenue will 25x.
The revenue they currently have and are projected to have in the mid term say they are massively overvalued (PPE), NFI why investors think their revenue will 25x.
GHG
Gold Member
And what do you think most of those alternatives are running in terms of hardware?
Sorry, the time to invest was last year when it was heading towards it's lows.
If you want to invest for the long haul at these multiples then good luck, you'll need it.
James Sawyer Ford
Member
AI isn’t going away. NVIDIA’s valuation will though
True, like image processing. The last company I worked at used AI via PyTorch to process images with no human intervention. They used Nvidia cards. I also interviewed at a health science company that was using image processing to look at pathology slides to look for and identify cancers. This field is just looking for uses.
newtypepilot
Member
E-Cat
Member
Will it if they’re the #1 global supplier of compute to train these AI models?
James Sawyer Ford
Member
Maybe they will be, but more likely they’ll face significant competition
Chips are used in everything already. No reason why they should be valued at current levels just because of the types of chips used give them a short term advantage
E-Cat
Member
Even a one-year lead would be a significant advantage. Not to mention the volumes enabling better pricing. A couple years back, I thought ASICS like the Google TPU would mean the big players would stop using NVIDIA's GPUs for training. But clearly I was wrong, OpenAI still uses them, Google and Meta will use H100s, Tesla too even though they're working on their DOJO.
Time will tell what happens, but we're soon to the point where one or two years could be the difference between GPT-x and AGI. NVIDIA has the momentum, the ecosystem, the API, the network... a lot of inertia, very little time to catch up.
GHG
Gold Member
Not sure why you would limit yourself to effectively cloud streaming from an Xbox Series X (at that point you may as well purchase the console) but you do you.
StereoVsn
Member
Yeah, on one hand I am happy Nvidia, MS and Apple have been doing well, on the other hand...
James Sawyer Ford
Member
We've seen this story before with many companies developing their own chips in-house. Apple designs their own chips. The specialized AI chips NVIDIA develops won't be a massive first mover advantage, and eventually cheaper entrants will create their own custom chips for their specific purpose. AI won't be solved in a year with the entire infrastructure fully developed. It's an ever evolving concept.
NVIDIA is priced to basically win the entire future of the AI industry from now into perpetuity and maintain massive margins and capacity while doing so. NVIDIA gets a price for their chips at the point of sale. Actual AI software companies get ongoing SaaS revenue streams. There's no better term for it than "speculative bubble".
CISCO in 2000 traded at 10x sales and had fantastic earnings growth for the last two decades as they "built the infrastructure of the internet", but investors that paid a high 10x sales price in 2000 still haven't recovered their initial investment.
Last edited:
You kinda missed the boat for big gains ... but I'm sure over time they will keep going up.
SF Kosmo
Al Jazeera Special Reporter
Yeah probably. nVidia rode the crypto bubble hard and then it popped, something which caught them a bit off guard with their high end card pricing this gen.
But AI isn't going to collapse in the same way. It might be riding a bit of a bubble as people scramble to add AI functionality to everything, but it's also a technology that is here to stay and that people will care about going forward. nVidia is way out in front in this space for both consumer and commercial implementations.
Puscifer
Member
Pretty much, these cards are for workstations and data centers, not gaming. There's cheap ai processing to be had in bulk than trying to sell workstation cards exclusively anymore.
These days I'm more of a reader and outdoors but it's looking more and more likely a PS5 and my current PC are the end of the line for gaming for me.
5-10 year dev times for console games, Increasing prices for PC hardware, the lack of ownership on the console side and the sabotaging of physical media, everything feeling more and more like a store front than a full product (waiting on street fighter 6, fuck what they did with 5) just starting to feel like everything's becoming a neoliberalhellscape where anything that can be sold, will be sold.
Also because I know it'll come up. Cost isn't the concern, but if I make 100K a year a high end PC went from 1K to 2-3K and the idea of putting 2-3% of my yearly income is fucking nuts! I only have a 4080 because I let go of a retro collection that felt like it was owning me (I'm not some YouTuber with 50 book cases of shit) and made a great chunk of change so it was worth it.
Joint venture with Apple and Intel to push some open ML language.
AMD 7xxx series is already very competitive in ML benchmarks against RTX 4xxx
SmokSmog
Member
Jensen:
