Nvidia shows signs in [2023]

[neckthrough]

NVidia on their part are happy to ship less consumer GPUs as demand cools due to high pricing since they are getting higher margins. Revenue and profits are up while GPU shipments are down. That's fine for them as a company. It only hurts them if a competitor can come in to take advantage of the the fact that NV are willing to sell less GPUs to less customers are a higher price.

I don't know where Nvidia's headspace is, but in general corporations *do* understand that consumers are more sentimental than businesses. This is especially true with today's influencer culture that (for better or worse) enjoys serious sway over consumer sentiment. Therefore, even though math may dictate the optimal (price*volume) point to be at a higher price point in the short term, this may end up being sub-optimal in the long term if it drives consumers away from the brand due to negative sentiment.

I guess the other potential problem with going with lower shipment/higher pricing (margin) strategy is that you reduce the amount of GPUs in the PC gaming space with new tech which reduces the incentive for developers to try to fully take advantage of those newer features while also pushing them towards console first programming WRT to features and tech.

Yes, this too. It's similar to the consumer-sentiment problem in the sense that maximizing short-term profits may come at the risk of longer-term profits.

But that's a longer term problem that NV might be willing to push into the future as long as they can continue to maintain high margins and enough people are willing to pay those high prices.

It's certainly possible, but NV doesn't appear to be a short-sighted company. A rational thesis would be that they are aware of the risk and are balancing it.

All that said, there's a difference between maintaining vs. lowering vs. increasing margins. Based on what I know about the state of the foundry industry I believe they are simply choosing to maintain overall margins on the consumer GPU side, and the cost increases are being transferred to the consumer instead of being absorbed. One may still ask whether it's prudent of them to do so, and whether it would have been better for them to eat some of the costs to keep the market more vibrant (the same short-term vs. long-term argument).

 

[trinibwoy]

Oh no, it is ruined for everybody. People are paying higher prices, but it doesn't mean they're happy about it. They still get hurt from this. As does everybody. Most people are just too selfish or short-sighted to understand the damage they do by giving in and cementing these higher prices forever.

Hurt in what way? Your premise doesn’t really make sense. People don’t willingly spend money on things that make them unhappy. If you choose to spend the money clearly it’s because you’re agreeing to the price!

If we take your argument to its logical conclusion then nobody is happy buying anything that’s not free.

 

[trinibwoy]

One may still ask whether it's prudent of them to do so, and whether it would have been better for them to eat some of the costs to keep the market more vibrant (the same short-term vs. long-term argument).

The IHVs obviously have immensely more information than we do on how much product is moving at various price levels. I think the 4080 launch price was stupid but some bean counter at Nvidia decided it would sell sufficient quantities at that price. We can only guess whether they were right.

In terms of sentiment, the best gauge is sales. If Nvidia’s market share isn’t dropping then clearly sentiment isn’t a problem. The idea that people are pissed at Nvidia but still throw money at them is a bit silly.

There is a huge difference between venting on an Internet forum (empty words) and actually shelling out hard cash. The latter reflects how you actually feel.

 

[techuse]

You should have also included out-of-stock items.

It doesn't change anything.

 

[DavidGraham]

Based on what I know about the state of the foundry industry I believe they are simply choosing to maintain overall margins on the consumer GPU side, and the cost increases are being transferred to the consumer instead of being absorbed

NVIDIA also burns lots and lots of R&D cash on software features (DLSS/Reflex/AI) and game graphics implementations (Ray Tracing/Path Tracing/Remix, etc), they constantly invent new graphics algorithms or train upscalers/denoising models. As an enthusiast I appreciate this, as it helps PC gaming set itself apart from consoles, I wouldn't mind paying an NVIDIA tax for this, as no other manufacturer seems interested in doing so.

 

[arandomguy]

The IHVs obviously have immensely more information than we do on how much product is moving at various price levels. I think the 4080 launch price was stupid but some bean counter at Nvidia decided it would sell sufficient quantities at that price. We can only guess whether they were right.

In terms of sentiment, the best gauge is sales. If Nvidia’s market share isn’t dropping then clearly sentiment isn’t a problem. The idea that people are pissed at Nvidia but still throw money at them is a bit silly.

There is a huge difference between venting on an Internet forum (empty words) and actually shelling out hard cash. The latter reflects how you actually feel.

Yes but to be fair they also initially thought the RTX 4080 12GB would be fine at $900 and then retracted it. That's not the only example of an IHV essentially making what you would classify a mistake so the market research and planning isn't exactly infallible.

Another aspect of this overall situation is I feel online enthusiasts seem to often forget that DIY retail is not the majority of discrete GPU sales and certainly not for Nvidia. Online enthusiasts themselves also aren't representative of all DIY retail purchases either.

For example in terms of sentiment the DIY enthusiast will likely look at components in isolation, so a $800 GPU should perform twice that of a $400 GPU, but what if it's a system purchase at $2000 vs $2400? It's only 20% more expensive. They'll also put more emphasis on gen on gen gains and comparatively to what they have in terms of numbers, but those aren't of importance to the vast majority of the market. Those people don't really have comparatives. The GPU and overall computer is just a device they use to them that the hardware itself and the numbers it can put up is not interesting to them.

 

[Deleted member 2197]

GeForce RTX 4090D for China to launch with reduced cores and higher base clock than RTX 4090 - VideoCardz.com

NVIDIA GeForce RTX 4090D is on track to launch next year The RTX 4090 Dragon for China reportedly features a cut-down AD102 GPU. NVIDIA is reportedly in the final stages of finalizing the alleged RTX 4090D, as indicated by renowned leaker MEGAsizeGPU, known for their accurate NVIDIA leaks. The...

videocardz.com

 

[digitalwanderer]

GeForce RTX 4090D for China to launch with reduced cores and higher base clock than RTX 4090 - VideoCardz.com

NVIDIA GeForce RTX 4090D is on track to launch next year The RTX 4090 Dragon for China reportedly features a cut-down AD102 GPU. NVIDIA is reportedly in the final stages of finalizing the alleged RTX 4090D, as indicated by renowned leaker MEGAsizeGPU, known for their accurate NVIDIA leaks. The...

videocardz.com

They ain't gonna be selling it in China, not if the US has any say in the matter:

Raimondo isn’t taking the announcement of these new chips lying down. “We cannot let China get these chips. Period,” she said, speaking at the Reagan National Defense Forum this weekend. According to Fortune (spotted by VideoCardz.com), she specifically called out “CEOs of chip companies…who were a little cranky with me when I did that because you’re losing revenue. Such is life. Protecting our national security matters more than short-term revenue.”


The Secretary had pointed words for Nvidia’s strategy of designing new variations of chips to slide under the TOPS power limit. “If you redesign a chip around a particular cut line that enables them to do AI, I’m going to control it the very next day.”

 

[troyan]

From AMD's MI300X presentation:

So H100 should be twice as fast as MI250X. Yeah, so muchfor theoretical performance numbers.

 

[Deleted member 2197]

NVIDIA Prepares to Launch RTX 5880 Ada Generation GPU for Professional Visualization

The precise specifications of the RTX 5880 Ada Generation are yet to be confirmed. However, in the context of NVIDIA's existing product range for professional visualization, the RTX 5000 Ada Generation targets the upper-mid market segment.

www.guru3d.com

This model, based on the "AD102" architecture, features a 256-bit ECC GDDR6 memory interface and offers 32 GB of memory. It includes 100 streaming multiprocessors (SM), surpassing the 80 SMs of the "AD103." Above the RTX 5000 Ada Generation lies the RTX 6000 Ada Generation, which is distinguished by its 48 GB of ECC GDDR6 memory and a comprehensive 384-bit memory bus, based on the "AD104" architecture. This model activates 142 out of 144 SM units. The forthcoming RTX 5880 Ada Generation is anticipated to fill the performance and capability gap between these models, likely providing an increased number of SMs, an expanded memory bus, and greater memory capacity.

 

[Deleted member 2197]

NVIDIA GeForce RTX 40 SUPER launch scheduled for January 17, 24 and 31 with RTX 4070 SUPER leading the way - VideoCardz.com

NVIDIA to split RTX 40 launch into three dates, RTX 4070 SUPER coming first, RTX 4080 is last ITHome have combined rumors from multiple sources which paint the picture when one should expect NVIDIA to unveil and launch the new RTX 40 SUPER series. Rumors suggest that NVIDIA is set to unveil and...

videocardz.com

 

[Deleted member 2197]

Dec 19, 2023

Adaptive Shells for Efficient Neural Radiance Field Rendering

Adaptive Shells for Efficient Neural Radiance Field Rendering

research.nvidia.com