Hacker News Reader

Comment by talldayo | Hacker News Reader

Businesses like OpenAI probably aren't long for this world. But who's going to unseat Nvidia? Seems like there's a strong core driving this rally with no competitors in sight.

View on HN

It's not about unseating. When NVDAs revenues are so concentrated from just a few customers, it only takes one to slow down and NVDAs multiple will contract.

https://finance.yahoo.com/news/single-customer-made-19-nvidi...

Isn't Nvidia overbooked though? If a few of their customers bail, who's to say that others won't take their place?

Sure, but overbooked leads to pricing power (there's also the question of who has billions to spend on GPUs). Stock price is about growth multiple, and NVDA is priced for perfection. Some say they are already priced for what's next after AI.

NVDA is making out like a bandit because they are selling the shovels (graphics cards) in this goldrush.

MSFT is hosting and investing in Open AI, but other than offering AI in Azure and O356 products, what is the business model and future for AI?

GOOG, META, and TSLA now have their own AI LLMs and chat gpts, but only GOOG has sort of a plan for AI, which is to replace search with it.

AMZN is the only competitor to Open AI and MSFT with Claude.

Again, what is the business model and plan for revenue?

I don't think google has a plan, I think its panicing.

Meta is sorta trying to make an AI assistant that lives on your glasses. I'm not sure they are going to make it practical before instgram dies.

I think this is a really poor analysis of both plans. Google and Meta have had AI-specific research labs for almost a decade now; TPUs, Llama and Tensorflow/Pytorch didn't materialize from nothing. They were part of a long-term, dedicated plan to explore AI capabilities. Both Google and Meta have Github accounts with thousands of open-source AI experiments.

You don't have to ascribe a business strategy to that any more than you ascribe a business strategy to whatever wack technology Apple patents next. These companies are using their R&D to test new technology, it's not hard to believe they do it for kicks since they already write-off hundreds of non-product developments a year.

You're making my point. If any of META, GOOG, AMZN, or MSFT decide maybe they can spend 20% less on GPUs, then NVDA will take a huge hit. It's not going to destroy NVDA or signal the end of LLMs, but will quickly put a damper on NVDAs multiple.

Or GOOG just uses their own chips: https://www.techradar.com/pro/google-is-rapidly-turning-into...

Exactly. Same with Meta.

hardware sales to datacenters is cyclical.

80% datacenter sales are to less than 8 customers.

marginal buyers - coreweave and xai are funded with speculative monies to the tune of $15B

so, expect pain.

How far behind is AMD?

https://www.reuters.com/technology/amd-launches-new-ai-chips...

Hardware wise, they have a shot. But software wise, they are relying on everyone else to start using something they can leverage.

I respect AMD for trying to get people to care about Open Source alternatives to CUDA, but the rest of the industry would rather fight Nvidia for scraps than work together to dethrone them. CUDA represents almost a decade of concentrated software design, and it's going to take a lot of work, proprietary or Open Source, to put it in it's grave.

I have not seen any explanation of why CUDA is so hard to replicate. Is it that hard... even with billions of dollars at stake no one can come up with an alternative?

> Is it that hard... even with billions of dollars at stake no one can come up with an alternative?

It's so hard that with trillions of dollars at stake, nobody can be bothered to compete. There are really two alternative routes that the industry can take to kill CUDA. I've mentioned it a few times in this thread but their choices are:

1) Revive OpenCL as a cross-platform GPGPU solution. Fund Khronos and get them to brush up the spec and make OpenCL into what Vulkan was for DirectX. There's a lot of work OpenCL needs before it catches up with CUDA, but a Free and Open Source, cross-platform GPGPU API would make a lot of Nvidia's software monopoly irrelevant.

2) Compete with proprietary, hardware-specific CUDA alternatives that implement only parts of CUDA's spec piecemeal. Basically entails trying to beat Nvidia at their own game, something the rest of the industry hasn't been very good at historically.

Option 1 makes the most sense; everyone wins if everyone's device speaks the same language. But the industry is too greedy to desire a true solution to CUDA's problem. OpenCL is derelict today because Apple, Microsoft and Google all feel as though they have better odds fighting CUDA alone. None of them are even close to Nvidia's performance profile or software capability, which is why I really feel that Option 1 is the only way forward.

I would love AMD in the race, but isn't it CUDA and the software written specifically for NVDA hardware that makes them dominant?

Question is not only about unseating, but also correction. Nvidia can continue making money, but be valued at more sensible level. Or there might be some trigger in scaling down manufacturing resulting the market cap dropping.

This may be Nvidia's "reasonable value" though. They have longstanding Nintendo contracts, significant presence in the desktop market, and they currently own the datacenter with no end in sight. Nvidia is shipping their own CPU designs for server hosting and reportedly has plans to make desktop CPUs as well. If optimistic AI-bros keep paying hand-over-fist for good compute (and I suspect they will), Nvidia will profit from the industry's incompetence.

and nvidia has the need for compute in general behind it, which i dont see going away anytime soon

Reply on news.ycombinator.com