Nvidia’s Requiring Big Change for RTX 5000!

[willhub]

On another note, anyone put together a spreadsheet and then onto a chart that displays the percentage of times Gamermeld and MLID have correctly, or incorrectly made predictions.

Would they have a higher level of accuracy than Baba Vanga?

 

[nomadd]

NVIDIA AI GPU Customers Reportedly Selling Off Excess Hardware​

The latest rumblings indicate that things have stabilized—so much so that some organizations are "offloading chips" as the supply crunch cools off. Apparently it is more cost-effective to rent AI processing services through cloud service providers (CSPs)—the big three being Amazon Web Services, Google Cloud, and Microsoft Azure.
The Information believes that some companies have already reduced their order counts, while others have hardware sitting around, completely unused. Maintenance complexity and costs are reportedly cited as a main factors in "offloading" unneeded equipment,

NVIDIA AI GPU Customers Reportedly Selling Off Excess Hardware

The NVIDIA H100 Tensor Core GPU was last year's hot item for HPC and AI industry segments—the largest purchasers were reported to have acquired up to 150,000 units each. Demand grew so much that lead times of 36 to 52 weeks became the norm for H100-based server equipment. The latest rumblings...

www.techpowerup.com

Maybe the 50 series GPUs will have to be more cost-effective now as the A.I. boom fades. Fingers crossed - this 1080Ti isn't going to last forever, Jensen.

 

[Grim5]

^^ not surprised

Last year there was articles saying companies in China were just trying to buy as much Nvidia as cards as they could, without having a solid plan how to use that investment, one large tech company was quoted as saying they are buying enough GPUs to last 10 years

So in that environment companies are going to overstock and buy too much. During the post Covid boom companies were trying to hire as much staff as possible and then found they hired too many and had to fire 20%, it's the same thing with GPUs

 

[nomadd]

..they are buying enough GPUs to last 10 years

Clearly preparing for the release of Star Citizen.

 

[Edward78]

Nvidia's Shady Behavior

⺠Thanks to Pulseway for sponsoring! Claim your free trial here: https://www.pulseway.com/enterprise/evaluation/?rfid=nca_influencer_yt_ufd⺠Check out today'...

www.youtube.com

Nvidia the GPU mafia.

 

[Nexus18]

If only the competition were able to deliver, nvidia wouldn't be able to dominate the market the way they do but reality is, said competitors can't compete. Personally I'm not going out of my way to buy a gpu that does not meet my needs/wants, if others do, well more power to them for sacrificing their experience. You can't put a price on your time and experience.

 

[Ayahuasca]

Nvidia's Shady Behavior

⺠Thanks to Pulseway for sponsoring! Claim your free trial here: https://www.pulseway.com/enterprise/evaluation/?rfid=nca_influencer_yt_ufd⺠Check out today'...

www.youtube.com

Nvidia the GPU mafia.

I'm a made man.

 

[Nexus18]

NVIDIA AI GPU Customers Reportedly Selling Off Excess Hardware​

The latest rumblings indicate that things have stabilized—so much so that some organizations are "offloading chips" as the supply crunch cools off. Apparently it is more cost-effective to rent AI processing services through cloud service providers (CSPs)—the big three being Amazon Web Services, Google Cloud, and Microsoft Azure.
The Information believes that some companies have already reduced their order counts, while others have hardware sitting around, completely unused. Maintenance complexity and costs are reportedly cited as a main factors in "offloading" unneeded equipment,

NVIDIA AI GPU Customers Reportedly Selling Off Excess Hardware

The NVIDIA H100 Tensor Core GPU was last year's hot item for HPC and AI industry segments—the largest purchasers were reported to have acquired up to 150,000 units each. Demand grew so much that lead times of 36 to 52 weeks became the norm for H100-based server equipment. The latest rumblings...

www.techpowerup.com

Maybe the 50 series GPUs will have to be more cost-effective now as the A.I. boom fades. Fingers crossed - this 1080Ti isn't going to last forever, Jensen.

And guess who is powering these cloud AI accelerated hardware solutions? Nvidia

We only work with the 3 main cloud providers and for any of our products using AI workloads, we're using nvidia gpu based VMs so whilst there won't be demand to buy directly from nvidia for said chips, they will still be getting profit from this one way or another.

We badly need another major player in the gpu sector now or/and some of the smaller ones need to partner up to take on nvidia.

 

[Edward78]

If only the competition were able to deliver, nvidia wouldn't be able to dominate the market the way they do but reality is, said competitors can't compete. Personally I'm not going out of my way to buy a gpu that does not meet my needs/wants, if others do, well more power to them for sacrificing their experience. You can't put a price on your time and experience.

I hope Battel Mage will be worth waiting for. Mostly it is probs. with drivers (Crazy Intel is have driver issues).

 

[Nexus18]

Nvidia's Shady Behavior

⺠Thanks to Pulseway for sponsoring! Claim your free trial here: https://www.pulseway.com/enterprise/evaluation/?rfid=nca_influencer_yt_ufd⺠Check out today'...

www.youtube.com

Nvidia the GPU mafia.

Just seen the source for this........

AMD GPU head accuses Nvidia of being a 'GPU cartel' in response to reports of retaliatory shipment delays​

That guy.....

No doubt there is probably some truth to it but take anything he says with a pinch of salt.

I hope Battel Mage will be worth waiting for. Mostly it is probs. with drivers (Crazy Intel is have driver issues).

Given how well intel have done, I'm quite hopeful for their next gpu. For their first dgpu, they did very well, on par with competing nvidia gpus for rt perf, upscaler nearly on par with dlss and they have sorted a lot of the main driver issues in a pretty quick manner.

 

[Grim5]

Herkel gonna Sherkel

 

[nomadd]

And guess who is powering these cloud AI accelerated hardware solutions? Nvidia

We only work with the 3 main cloud providers and for any of our products using AI workloads, we're using nvidia gpu based VMs so whilst there won't be demand to buy directly from nvidia for said chips, they will still be getting profit from this one way or another.

We badly need another major player in the gpu sector now or/and some of the smaller ones need to partner up to take on nvidia.

Microsoft, Meta, OpenAI, Qualcomm, Samsung, Google and several others are already designing their own AI NPUs. Nvidia have a lead, but the big players all want to go their own way and not be tied to Nvidia. I mean, just look what's happened to graphics cards..

I can see, after a few years of in-fighting, the big players merging designs around some common standard, without the IP lock-in that ARM has on the mobile space. The EU already has a consortium underway, for example. If A.I. really does have staying power, then no government is going to let one single player dominate.

 

[Nexus18]

Microsoft, Meta, OpenAI, Qualcomm, Samsung, Google and several others are already designing their own AI NPUs. Nvidia have a lead, but the big players all want to go their own way and not be tied to Nvidia. I mean, just look what's happened to graphics cards..

I can see, after a few years of in-fighting, the big players merging designs around some common standard, without the IP lock-in that ARM has on the mobile space. The EU already has a consortium underway, for example. If A.I. really does have staying power, then no government is going to let one single player dominate.

Yup of course they will want to do that and I can't blame them, it will pay of long term for them to self invest. Problem is Nvidia are so far ahead and now got huge capital behind them to potentially get even further ahead.

 

[Gerard]

NVIDIA AI GPU Customers Reportedly Selling Off Excess Hardware​

The latest rumblings indicate that things have stabilized—so much so that some organizations are "offloading chips" as the supply crunch cools off. Apparently it is more cost-effective to rent AI processing services through cloud service providers (CSPs)—the big three being Amazon Web Services, Google Cloud, and Microsoft Azure.
The Information believes that some companies have already reduced their order counts, while others have hardware sitting around, completely unused. Maintenance complexity and costs are reportedly cited as a main factors in "offloading" unneeded equipment,

NVIDIA AI GPU Customers Reportedly Selling Off Excess Hardware

The NVIDIA H100 Tensor Core GPU was last year's hot item for HPC and AI industry segments—the largest purchasers were reported to have acquired up to 150,000 units each. Demand grew so much that lead times of 36 to 52 weeks became the norm for H100-based server equipment. The latest rumblings...

www.techpowerup.com

Maybe the 50 series GPUs will have to be more cost-effective now as the A.I. boom fades. Fingers crossed - this 1080Ti isn't going to last forever, Jensen.

That's his new leather jacket order cancelled with that news.

 

[Edward78]

I have a RTX 3060ti 8 gig now, if I can play games like Mortal Online 2 with no issues, I am fine with it.

 

[Kelt]

I have a RTX 3060ti 8 gig now, if I can play games like Mortal Online 2 with no issues, I am fine with it.

Well, that's always the case when upgrading a CPU or GPU, does it do what you require.

If yes, don't upgrade, if no, then upgrade.

 

[Joxeon]

Yup of course they will want to do that and I can't blame them, it will pay of long term for them to self invest. Problem is Nvidia are so far ahead and now got huge capital behind them to potentially get even further ahead.

Nvidia’s done well because the GPUs work on a wide variety of things, like how GPUs were good for mining BTC at the start but then once ASICs showed up they quickly got left behind and the same will probably happen with AI as companies design their own ASICs for the specific tasks they need doing.

 

[Dicehunter]

I hope Battel Mage will be worth waiting for. Mostly it is probs. with drivers (Crazy Intel is have driver issues).

For the first time ever I am really cheering for Intel to do very well with their GPU's. If they can become competitive at the high end and keep prices sane then I'll make the switch in a year or 2.

 

[Edward78]

For the first time ever I am really cheering for Intel to do very well with their GPU's. If they can become competitive at the high end and keep prices sane then I'll make the switch in a year or 2.

Yes, I have big hopes & MS is coming out with DirectSR. That will do upscalling with Windows, so all GPUs will be able to do that.If mafia Nvidia doesn't stop it, but it is MS not some small company doing DirectSR.

 

[Nexus18]

Yes, I have big hopes & MS is coming out with DirectSR. That will do upscalling with Windows, so all GPUs will be able to do that.If mafia Nvidia doesn't stop it, but it is MS not some small company doing DirectSR.

Microsoft have been working with amd and nvidia on directsr and it's not another upscaling solution by them, instead if it is an api that will be embedded to allow all upscaling solutions like xess, dlss and fsr to be integrated into games.