NVIDIA is valued far in excess of potential GPU sales. They need the hype train. Running afoul of the US government is bad for the hype train.

This just isn't true. NVDA's price/earnings ratio is 22, which is less than MSFT.

Their P/E ratio is lower than in 2020 (before AI got rolling, and way below their 2023 peak.

If anything - based on current sales alone - NVDA is undervalued.

I’m not sure that Hardware has the same moat as Service/Platform.

If (it’s a big if), NVDA get leapfrogged in hardware capability it will be trivial for the next generation of deployments to stop using them entirely (Google and Microsoft are already developing custom inference chips, as are others)

I don't think this is true. AMD's hardware, on paper at least, offers the same or better performance than Nvidia. Eg the Mi355x vs B300 is close. The former released before Nvidia's GPU.

The same story goes back for years and years in consumer hardware too. AMD GPUs often had more raw compute capability, but fell short in actual game performance (or other applications).

Nvidia's moat is, at least partially, the software to make the hardware useful.

This is a great point I overlooked, thanks.

The hyperscalers are one story but research/consumer/“Everyman” inference is likely very different.

>NVDA is undervalued

i would like to, but won't/can't buy their overpriced crap -> they're losing a customer -> they're doing something wrong -> their valuation is just about 0 to me, ballpark estimate

Obviously he talks about stocks

Nvidia is king of circular financing. peak accrual earnings include non-cash items and demand Nvidia is funding itself. On a cash (real actual reality) basis the multiple is 33 to 50

the 5T is crazy to me.

I wish for US to win the Chinese race if there's one. I would agree that the Chinese getting access to GPUs, building models, making them open weight and also sharing has really pushed the American labs. Without Chinese models, we would probably be 1 year behind where we currently are. Competition is a healthy thing. This competition has increased intelligence so fast and created ridiculous demand on hardware. So it's been great for Nvidia and their stock. I don't know what Jensen knows nor do I believe he could have anticipated how the field would play out. Whatever happens, it's been one heckuva ride for him. Unfortunately there are many people that would love to get into his position, the pay is going to be delicious...

These are strong and unsubstantiated claims.

The main competitions US labs are facing, are from one another at this point, not from open weight models. If anything, the competition from new entrants in the west (Meta, xAI, even Google if you think about it) has pushed things even faster, but regardless, I don’t see how OpenAI and Anthropic would have been any slower if there were no open weight models.

In my opinion it is the open weights that are pushing as much in tandem. Without them it would be easier between the top labs. They must have a large enough gap between OW to justify that extra cost. A business owner may prefer to be happy with 8/10 AI output rather than spend 5x more for that extra 2 points

Given the risks of AI tech, one could argue quite easily that "we would probably be 1 year behind where we currently are" would be a better outcome.

I wish this race to the bottom would stop.

Why do you wish for US to win? I think it's a quite scary moment if I, to be able to work, need to pay a hefty tax in tokens to a few big American corporations instead of having a server in my cellar with a few GPUs I control or having a healthy ecosystem of competition serving the models we need daily.

In the long term, the strategic interest of Nvidia is to avoid having to deal with a business entity that is too powerful.

While yes, they also have interest in open weight, it does increase their market after all, I'd be surprised if that was on the same level, given they are extending mass credits to US frontier labs and neoclouds on condition that they'd purchase their GPUs and even provide revenue guarantees.

That is something they simply cannot do with Chinese entities. Allegedly smuggling GPUs may be feasible, but that'd probably ruffle to many feathers in both the US and Mainland China, especially as we know they will pop a bubble to their own (temporary) detriment if it gets out of control as seen with Evergrande a few years ago.

> Chinese labs releasing models that people can run themselves on Nvidia GPUs creates more potential buyers of Nvidia GPUs

Nvidia would rather sell those big interconnected gpu racks than individual GPUs.

> I don't think it's good for Nvidia if everyone has to go through a few labs APIs to get access to AI inference.

Maybe but isn't a business customer better for Nvidia than regular customers? They buy in bulk and commit to multi year purchases. Bonus points for circular deals

I'm pretty sure they would prefer to have a diversified line of products with customers willing to buy them. Should apple just stop selling everything except their most expensive iPhone, because it's expensive and popular?

Would that not perhaps hurt their product ecosystem? What if people decide they don't want so much of one product, and would actually like to buy some of those cancelled products at some point in the future? I guess they'd have to go to a different company then!

Who said anything about individual GPUs or consumers?