NVIDIA to Consume 77% of Silicon Wafers Dedicated to AI Accelerators in 2025

[AleksandarK]

Investment bank Morgan Stanley has estimated that an astonishing 77% of all globally produced silicon wafers dedicated to AI accelerators will be consumed by none other than NVIDIA. Often, investment research by large investment banks like Morgan Stanley includes information from the semiconductor supply chain, which is constantly expanding to meet NVIDIA's demands. When looking at wafer volume for AI accelerators, it is estimated that in 2024, NVIDIA captured nearly 51% of wafer consumption for its chips, more than half of all demand. With NVIDIA's volume projected to grow to 77%, this represents more than a 50% year-over-year increase, which is incredible for a company of NVIDIA's size. Right now, NVIDIA is phasing out its H100 accelerators in favor of Blackwell 100/200 and the upcoming 300 series of GPUs paired with Grace CPUs.

NVIDIA is accelerating its product deployment timeline and investing a lot in its internal research and development. Morgan Stanley also projects that NVIDIA will invest almost $16 billion in its R&D budget, enough to endure four to five years of development cycles running three design teams sequentially and still delivering new products on an 18-24 month cadence. The scale of this efficiency and development rivals everyone else in the industry. With all this praise, NVIDIA's Q4 revenue report is coming in exactly a week on February 26, so we have to see what its CEO, Jensen Huang, will deliver and show some estimates for the coming months.



View at TechPowerUp Main Site | Source

 

[noel_fs]

i have a feeling this is soome bullshit to bait people into buying stocks

 

[AnarchoPrimitiv]

In a sane world, this company would be broken up.

 

[Space Lynx]

RIP world's energy grid. so much for the 9 watt lightbulb that mimics a 75w lightbulb in brightness changing the world. ah, what a dream it once was.

 

[R0H1T]

In a saner world there'd be people doing a lot more with their time/money/power than invest in Crypto as well but then here we are

 

[mb194dc]

Imagine how much waste there is going to be from all this insanity. Who's paying for it all on the front end, you can invest hundreds of billions in hardware, actually need to charge some end user to recoup the cost. Most of it looks like dead money.

More so since it's been shown models as good as any can be created with a tiny portion of the hardware already in existence.

 

[JohH]

Well now you know why you can't buy a graphics card. Nvidia is going all in on DCGPU.

 

[FoulOnWhite]

Nvidia wants to be cyberdyne for real

 

[mechtech]

I would say that's pretty heavily reliant on 3rd parties!!

 

[Wirko]

Does anyone still build plain boring non-AI datacenters, the ones with plain old boring Epyc and Xeon (and Altra Ampere) chips inside, doing boring useful work?

 

[dont whant to set it"']

Say what one will, as trend goes, I have yet to be approached by an nVidia bribe; random internet's user out.

Those $1/2 Billions look chump's change by now after it's lawsuit by the EC(EU).
Intel got like half billion and AMD got like 1 whole billion as a result. Not quite Chad chance, a bit over, still: Profits first: safety maybe, what is safety? .

 

[ncrs]

Does anyone still build plain boring non-AI datacenters, the ones with plain old boring Epyc and Xeon (and Altra Ampere) chips inside, doing boring useful work?

Fortunately NVIDIA's AI chips (and GPUs) can be used for other types of computations, and often significantly outperform CPUs in scientific computing.
So a theoretical AI bubble burst is not going to leave a ton of "useless" hardware.

 

[Vayra86]

RIP world's energy grid. so much for the 9 watt lightbulb that mimics a 75w lightbulb in brightness changing the world. ah, what a dream it once was.

We already killed that dream ourselves by adding a few dozen extra lightbulbs to the average house In quite the same way there are now multiple screens in a single house where we used to have just a TV;.

Just like we're killing it now with crypto, AI and other brute force calculations with questionable gains.

All it takes is a look in the mirror...

 

[Luminescent]

Governments are throwing money at Nvidia to research AI.
Basically YOU, tax payer, you are paying with your taxes so they can research AI to a point so they can replace you and fire you.

 

[Mr. Perfect]

Is that picture of Jensen photoshopped, or did he really walk across the stage holding a giant die shot like a shield?

 

[DaemonForce]

Is that picture of Jensen photoshopped, or did he really walk across the stage holding a giant die shot like a shield?

If you were around to watch CES you would know it's real. The world we live in, this is it.