[nickpeterson]

People keep yelling about nvidia stock but that feels like a huge bubble. AI disillusionment will hit and the stock will implode. Nvidia hasn’t made any inroads on producing actual systems, just gpus. Once Apple or Microsoft have a fast enough chip (TOPs wise), nobody will care about nvidia lead except in the datacenter. Seems like a failing position to me.

[eitally]

I think I'm not adequately understanding your comment. Nvidia has huge software teams building accelerators that optimize application of their GPUs for all kinds of applications. They also now offer their own "cloud" and have partnered with the hyperscalers for integration as well as GTM. Those same hyperscalers are also almost unilaterally driving Nvidia's growth the past few years. The cloud market is still growing and will continue to do so. Nvidia's consumer business almost doesn't matter anymore.

[bob1029]

I think much of this boils down to perspectives some of us have regarding the value of in-house manufacturing capabilities relative to design time & software capabilities.

I get arguments that maybe one fab is better than the other, but what about all of them combined? All of our modern chipmaking capability all at once.

Nvidia has no factories. You can ship their output on a USB flash drive. Valuation: ~3.1T.

Intel, TSMC and Samsung have all the factories. Every modern chip made on earth in this circle. Combined valuation: ~1.1T

This is simple napkin math for this arbitrary retail investor. I don't know when the music will stop but it absolutely will.

[wormlord]

I agree with this sentiment overall, but we have to remember that for valuations, profit and growth is really all that matters. Even if an industry is irreplacable.

I think I read this from Warren Buffett, but basically as of the early 2000s the airlines, in their entire history, had only managed to break even. If you had bought an airline company in 1940 and held it until 2000, you would have never profited from it. The business itself would be worth significantly more, but your only exit strategy would be to sell.

I haven't looked at any of these company's balance sheets, but it might be that semiconductor fabbing is less profitable and has less room for growth. In the short term that's all that matters. The question is if Nvidia can hold on to its current growth and margins (I don't think it can).

[matwood]

Are you saying a software company has less value for some reason? See MS. Or are you saying a hardware company that doesn't own their production has less value? See Apple.

I think we've seen over many years at this point that there is huge value in the final product.

[_zoltan_]

hard depreciating assets (factories) + selling chips are way worse than actually selling an ecosystem. NVDA is just getting started :-)

[CoastalCoder]

I've worked on AI-hardware software stacks for several well known companies.

It's impossible to overstate to advantages that CUDA, it's documentation, toolchain, and nSight software provide to outside developers.

The closest thing I've seen to nSight Systems software is Intel's VTune. But that's just one piece I'm a much larger puzzle, and last I checked, VTune was only for Intel CPU.

AFAICT, Nvidia's software seriously reduces the ramp-up time for new developers to write kernels or apps that make good use of the available hardware.

E.g., nsys-ui (like VTune) recognizes anomalous profile results, and makes solid suggestions for next steps. I don't know of other software that does this (well), although maybe I'm just uninformed.

[Jlagreen]

This is wrong, please check what DGX is.

DGX is a complete data center from Nvidia where Nvidia is the supplier of everything themselves:

- CPU+GPU from Nvidia - Rack from Nvidia - Interconnects + networking from Nvidia - SW from OS to application framework from Nvidia

The only thing Nvidia really needs partners with DGX is memory (RAM + SSD).

One reason Nvidia's margins are so high is because they provide the whole data center so while competition has to split margins (AMD/Intel + SMCI/DELL + Broadcom/Arista + Cray/HPE).

[_zoltan_]

no, this is not true. DGX is (right now) a 4U box with 8xH100 on a daughterboard.

Don't believe me? Do you believe nvda?

https://www.nvidia.com/en-sg/data-center/dgx-h100/

[edward28]

There's also the dgx nv72

[kcb]

DGX is the branding for the whole platform too. It's right on that page...

[redleader55]

Nvidia doesn't only produce GPUs. They have their hands in all sorts of interesting technologies from high speed, performance NICs to switches. The GPUs in the data centers don't just function as an individual unit, but they are able to use the "backend" NICs of the server, in conjunction with their proprietary NCCL library to send data with as close as possible to "zero-copy" from GPU to GPU with very good horizontal scalability. As you can imagine the network fabric behind this is quite important, and Nvidia having one the best Infiniband switches helps keep this monoploy.

There are many companies working on alternatives at the moment, but it will be a while until Nvidia can be replaced.

[htrp]

>Once Apple or Microsoft have a fast enough chip (TOPs wise), nobody will care about nvidia lead except in the datacenter.

thats gonna take a while

[imtringued]

Google has powerful TPUs. The real question is, why isn't Meta building their own?

[wmf]

Meta is on their Nth generation.