I don't understand ai valuations right now.
NVDA has market cap of 1.8T and all they're doing is the middle part of the value chain.
OpenAI is doing all the actual productization and they get a valuation of 20x less?
NVDA has market cap of 1.8T and all they're doing is the middle part of the value chain.
OpenAI is doing all the actual productization and they get a valuation of 20x less?
With that said, I do believe Nvidia is overvalued - if it triples profit its PE ratio would still be 30 (i.e. 30 years to return on investment), while there's a fairly good chance someone would catch up to them in 10-20 years.
Using trailing PE shows an inaccurate picture for a high growth company so it makes more sense to just take the last quarter and project forwards.
The next thing to explain the discrepancy between Nvidia's valuation and OpenAI's would be that Nvidia's monopoly position effective eats into the profits of the AI startups for the foreseeable future. Had OpenAI already been profitable, its valuation would have exceeded 86B.
If you are willing to just take quarterly revenue which I think is reasonable for NVidia, it is valued at around 40 times the current estimated earnings for this quarter which isn't too overvalued.
The bigger thing I worry about NVidia is not current earnings but the possibility that the earning won't last when either AI wave fades off or competitors enter the market leading to loss in margin.
There’s more hardware than nvidia but it takes some time to unwrap the proverbial matmul from CUDA. Much more hardware is in the pipeline, it just (ahem) has to be much better than nvidia to make migration from CUDA worth it. groq is one such limited but extremely impressive example IMHO.
I think that's the bet. Maybe someone could catch up, but what are the chances incumbents buy new entrants or otherwise defeat them? What's the chance regulators will get involved? If you rate those favorably for NVIDIA, you'll price it higher.
I disagree. Google have their tensor chips – whether they work well enough for those outside of Google is somewhat irrelevant, they clearly work for Google who are going to be one of the major players in serving AI for the forseeable future (I'm biased, but seems clear to me). Microsoft have their own chips on the way, rumoured to be this year. Amazon have their own chips on the way, rumoured to have been in development for a number of years.
All 3 of these companies sell Nvidia hardware on their cloud offerings because it's what buyers want, but with Nvidia's pricing and resource constraints there is a huge pressure on the cloud providers to push customers into their own offerings. I don't expect any to stop offering Nvidia chips directly, but I'd bet that in the next few years all of their hosted value-add services will be on alternatives (i.e. hosted AI, inference, training, etc where customers don't need to know the hardware).
Nvidia have more of a moat than OpenAI for sure, but I think Nvidia's best days are 2023/4, and that things will look very different soon.
Similarly when IBM manage to start producing volumes of their neuromorphic chip, Google’s TPUs, Nvidia and even groq LPUs may seem obsolete.
I think this is most likely a temporary blip. GPUs were a bit of a commodity 5 years ago, with Nvidia, AMD, and Intel all producing reasonable stuff. Large AI accelerator chips weren't much of a market ~5 years ago, Nvidia were first to take the market, but in a few years time they'll also be back to commodity status.
Nvidia have a small moat with CUDA, but their eye-watering prices are a huge incentive for users to try alternatives, and ultimately the current price is built on them being the only provider of GPUs with 40/80GB of memory. That's the fundamental enabling technology, and that's not particularly tricky for competitors to replicate.
Nvidia may be the "best" AI accelerator chips on the market for years to come, but being 20% better and 20% more expensive than AMD, and all the cloud providers using their own in-house chips where they can, is not a $1.8Tn company as far as I can tell, it's much more like what Nvidia were ~5 years ago.
so question is: where are these guys' ML chips? sorry but AVX512 is not something that provides enough juice, and apparently some smart-head at Intel decided to lock end-users out of it?
because, honestly, it was not NVidia who kicked the GPUs forward, but these brave CUDA devs who actually created some valuable software to run on top of them - first for crypto mining, then for the LLMs and NNs in general.
honestly - i start to really despise this company, even though there's a 3090TI in my home box. and with the most recent talk given by CEO - fingers crossed someone comes and eats their lunch, they so much deserve it.
NVIDIA make bank for every single failed AI venture, if they buy their chips.
If Nvidia can truly print money as long as TSMC can make chips, as claimed above, I'd argue that's at least in a gray area for commodities. The chip is effectively a raw material as far as Nvidia is concerned, and it acts much like gold or oil in that scenario as the argument is that there is an unlimited market willing to buy any GPU that Nvidia can create.
Having a monopoly doesn't mean the product itself isn't a commodity. De Beers has a monopoly on diamonds, but I'd expect diamonds to fit into most people's definition of a commodity.
That's the definition everyone in economics uses and NVIDIA has never fit that definition. NVIDIA GPUs are not fungible. Intel and ATI GPUs are not replacements as far as AI is concerned. There are zero substitutions. They are not commodities, they are differentiated goods that no one else can produce.
> Having a monopoly doesn't mean the product itself isn't a commodity. De Beers has a monopoly on diamonds, but I'd expect diamonds to fit into most people's definition of a commodity.
Yes because the definition of a commodity is fungibility not the number of vendors. You can go online to Alibaba now and buy a bag of synthetic diamonds that are perfect replacement for mined diamonds. De Beers never really had a monopoly, they just controlled the market for consumer diamonds until the 80s.
NVIDIA has a monopoly because its product is not a commodity.
NVIDIA’s products can colloquially be described as commodities in some contexts, as GPUs from NVIDIA can be interchangeable with GPUs from other manufacturers like AMD. For specific tasks like gaming or basic computing, the brand may not matter as much as the specifications, making them somewhat commodity-like in those scenarios
Whilst this isn’t the traditional economic definition of commodity, speaking loosely, I think it’s fair enough to describe GPUs as a form of commodity. The important thing is what’s being communicated, not the semantic definition. The above comment’s point was pretty clear IMO
For AI workloads I'm pretty sure they aren't, which is why everyone is trying to buy NVIDIA GPUs.
If they were a commodity there would not be such intense competition for NVIDIA GPU allocations, because they would be easily interchangeable with GPUs from another source.
But the argument is, that Nvidia should follow a standard valuation model, since their product is a commodity.
Those claims are both incorrect. This either rests on a misunderstanding of what a commodity is, or on a misunderstanding of Nvidias Position in the AI segment.
As it stands at the moment, they are not a commodity in this field, they can not be replaced, and thus you cannot apply a more standard valuation model.
Nvidia is not valued as a commodity because now and in near future people expect them to keep their superiority that allows them to charge 30000+ for one card.
Nvidia P/E is 95.
OpenAI's valuation is 4 times more insane.
OpenAI P/E is 415
Nvidia P/E is 95.
OpenAI is valued over 4 times Nvidia based on revenue.
What about AMD (MI3000), Google (TPU - actually designed by Broadcom).
It's really TSMC (who actually make NVIDIA's chips, as well as those for AMD, Google, Apple ...) who are close to making things that no-one else can, although afaik Samsung are close.
While what they sell isn’t intelligent, it has utility and that’s worth something. Since their product seems to be at the forefront and they’re really the trendsetter with others playing catchup, they can command a premium.
The eventual winner will almost surely be worth more than Nvidia when the rubble clears, but they, and all the losers and midfielders won't get to that day without forking over billions of their dollars to....
OpenAI has no definitive control on the service they provide, eg any one that can provide more value or cheaper on Nvidia hardware is a direct competitor that may be able to afford to pay more for hardware, limiting OpenAI access to its critical ressource.
Hence nvidia is the elephant that structurally takes the money, the ones that make the safest big returns on long term.
OpenAI wants to create machines that are more capable at almost all economically valuable labor than humans. Such technology would be incredibly valuable.
If it seems likely they’ll achieve it and actually return a significant part of the created value to investors, valuations that are crazy by old metrics will follow.
I would be curious when an educated-on-the-matter person's opinion was best guess on when NVIDIA's competition will catch up
oh, and don't forget Graphcore (also rumored to be bought soon): https://www.tomshardware.com/tech-industry/artificial-intell...
NVDA may yet see some competition, as we drop our complexity standards and embrace a one-size-fits-all general MLLM architecture.