This video analyzes the RTX 50 series GPU launch, highlighting several reasons why the speaker advises against purchasing them. The speaker covers issues related to the launch, pricing, stock availability, hardware problems, and software issues, ultimately comparing the situation to competitors AMD and Intel.
Firstly, don’t worry. The title is not clickbait. Lately, the GPU market has gotten hectic, especially on Nvidia’s side. That's why I need to fill you in. Ready? Let’s dive in.
I'll start with Nvidia since it's our main topic and the company that triggered this whole mess. The RTX 50 series had an absolutely catastrophic launch. The launch, the pricing, the stock availability, the hardware and software issues — all of these were more than enough to frustrate users.
-The Launch Let’s start with the launch. During the launch, they made highly inconsistent promises. They said “The RTX 5070 will be as powerful as the RTX 4090”. However, in terms of performance, the RTX 5070 couldn’t even outperform the 4070 Ti. In fact, it only manages to perform similarly to the 4070 Super. What a joke. This was all a marketing strategy. The RTX 50 series can distinguish itself solely with its Multi Frame Generation feature. As it’s a long name, I’ll just shorten it to MFG. This MFG technology generates artificial frames with the assistance of AI. Thanks to this, your FPS can increase by up to 2 to 3 times; though this results in degraded image quality and increased input lag. By the way, I don’t think that we are talking about this enough. We can get higher FPS thanks to the MFG technology, that’s correct but the image artifacts are far from negligible. Competitive games already don’t support these kinds of technologies due to these distortions and the input lag. So, do you really think the MFG technology is that important? I’m curious about your thoughts. Feel free to share them in the comments. So, what happens if we ignore the MFG feature? Besides the MFG feature, we can’t really see a noteworthy upgrade. Except the RTX 5090, no model significantly distinguishes itself from the RTX 40 series. Similar criticisms were made during the transition from the RTX 30 series to the RTX 40 series because despite the fact that RTX 40 series cards had lower power consumption, GPUs below the RTX 4070 didn’t offer a significant performance leap. But the RTX 50 series is in an even worse situation in this regard. All cards, except for the RTX 5090, don’t offer substantially different performance from the previous generation.
At this very point, we also need to talk about the RTX 5070 Ti. This card also offers a limited performance increase. Because of this, I want to name this card “RTX 4070 Ti-Super PLUS” instead. So you can probably imagine how the RTX 5070’s performance.
-The Price and Stock Moving onto the price and the stock availability, we used to be able to see nice products on the market despite bad product launches. But this time around, the products themselves couldn’t even make it to the market properly. Looking at the price graphic, you can see that all GPU’s except the RTX 5090 went down in price. Everything seems fine, right? At least that’s what we initially thought as well. But Nvidia shifted their production line to focus more on AI GPUs. And this led to GPU’s for the end user to be deprioritised. Considering how popular Nvidia is as well, it wasn’t hard to predict that inventory would deplete rapidly. Manufacturers such as MSI and ASUS typically sell cards for around 20 to 100 dollars more, which we consider normal. But due to stock shortages with this series, the graphics cards are being sold for almost two times the MSRP in some countries. RTX 5090, in particular, can reach exorbitant prices, like 4000 dollars. With these circumstances, even just buying an RTX 50 GPU series becomes extremely challenging.
-Hardware Issues Moving on to the hardware issues, as if there weren’t enough problems already, this time very serious issues have emerged with the GPU’s themselves. Furthermore, there isn’t just one issue; there are four critical problems.
The first and most significant issue is some GPUs having missing ROP units. This is, flat out, a manufacturing error. The count of ROP units, which are located in the core of the GPU, is directly proportional to the performance you will get. But in some GPUs, the ROP unit count is less than it should be. This means you don’t get the performance you’re supposed to from your card. For example, the performance of an RTX 5090 can drop to the performance of an RTX 4090… This is really unacceptable! Moreover, this issue is entirely down to luck, there is no guarantee you won’t face it as well. A user familiar with components can notice the difference and realise something’s wrong. But if they aren’t knowledgeable about components, they might not even realise it. Imagine saving up your money to buy an RTX 5080 graphics card, but it performs worse than it should, and you don’t even notice. If you’re using a GPU from the RTX 50 series, I recommend that you install the GPU-Z and check your ROP count immediately. If you have fewer ROP’s than you’re supposed to, you should absolutely request a replacement for your graphics card.
The second important issue is regarding the power cable. This problem is caused by the 12V-2x6 (Twelve volt two by six) power cable used on the RTX 40 and RTX 50 series. If this cable is not seated properly or if it sustains high power transmission for extended periods, an issue as severe as the power connector burning out can occur. Actually, Nvidia is not the only one at fault here. There is another organisation in charge of connector standards. As far as I know, this issue is more prevalent with the RTX 5090 and 5080 cards. So, if you are using these cards, I advise you to be extra cautious.
Our third issue is about the PhysX support. PhysX is a technology that allows for the addition of more realistic physics effects to games. This feature used to be activated by connecting an external device. Nvidia later bought the companies that developed these technologies and integrated them onto their GPU’s. So, for a while, there were dedicated PhysX units on GPUs. But as PhysX was rarely ever being used nowadays, Nvidia completely removed this support from RTX 50 series GPUs. But here’s an overlooked detail: Old games can still need PhysX support. For instance, if you enjoy older games and use a RTX 50 series GPU, and the game you want to play has PhysX features, you might be unable to play the game flawlessly even if you’re using an RTX 5090. The game might look different from how it should. You might need to take additional steps to get PhysX working. But I have good news as well: Nvidia made all of the source code for the PhysX system open source. This way, maybe this issue can be alleviated to some extent through community support.
Our fourth issue is the PCI Express 5.0 incompatibility. If you install an RTX 50 series card onto a motherboard supporting PCI Express 5.0, you might encounter compatibility issues. In such cases, you usually need to update the BIOS or change some settings. This issue can be particularly annoying especially with GPUs that need high data transfer speeds such as the RTX 5090.
It’s clear that Nvidia is going through a tough patch. But Nvidia is not the only player in the market; there are significant developments from AMD and Intel, which I want to touch on briefly.
AMD had previously announced that they wouldn’t compete in the high-end market with the RX 9000 series. However, AMD recently announced the RX 9070 XT and the RX 9070 GPUs as a powerful response to Nvidia after a considerable hiatus. These cards, whilst competing with the RTX 5070 Ti and the 5070, are in a much more advantageous position regarding price and stock. Plus, thanks to the improved Ray Tracing performance and Frame Generation support, AMD became a great competitor. On the software side, I haven’t seen them experiencing a serious issue. As the Wattman issue was resolved with the RX 7000 series, you no longer have to worry about that. AMD is also aware of this situation. They were a bit more relaxed with the RX 7000 series, but given the launch and sales success of the RX 9000 series, it’s almost certain they’ll intensify their marketing efforts.
Now let’s talk about the new player of the market, Intel. They made a great debut with their B570 and the B580 cards. The B570 is a bit redundant when the B580 already exists so I will use the B580 as our reference. In terms of performance, the B580 is comparable to the RTX 4060 Ti. It succeeds at having the highest price to performance ratio thanks to its technologies such as XeSS, high performance and the most important of all, having 12 GB VRAM. Its price ranges from 250 dollars to 350 dollars. Its power consumption is better than the previous generation; but the main feature that stands out is the cooling performance. A750’s Limited Edition and dual fan Sparkle cards operated at really high temperatures. But on the contrary to the A750, all cards of the B580 operate at pretty good temperatures. Besides these positives, there are some shortcomings on the software side. The drivers are continuously being updated but you can still encounter issues in some applications. However the main problem is suboptimal performance when paired with older CPUs. Specifically, when used with processors older than AMD’s Ryzen 5000 series and Intel's 12th generation, the B580 doesn’t deliver its expected performance. The actual reason for this is the lack of Resizable BAR support and driver issues.
Now, to summarise: Nvidia: Not doing well at the moment. AMD: Shaping up to be a very serious and powerful competitor. Intel: Continuing to heat up competition in the budget segment and steadily improving.
I’d like to add a few final thoughts on Nvidia. As you know, GPUs are extremely important in the field of artificial intelligence. The higher the CUDA core count, tensor core count, and memory capacity, the better the AI performance. Nvidia also supplies a significant portion of the graphics cards needed for the AI market. Nvidia made a lot of money thanks to their cards' high performance in AI related workloads, especially cards such as the H100 and the A100. Due to this, Nvidia has dedicated a very big portion of its production capacity to AI focused graphics cards, meaning the cards for the end consumer have taken a backseat. I believe many of the many problems with the RTX 50 series stem from this. It appears that Nvidia is not giving the consumer GPUs the attention they deserve, and this is causing problems for them. I am curious about your thoughts on this topic as well. Don’t forget to share your thoughts in the comments.
Alright, let's wrap up the video. If you liked the video, don't forget to hit the like button; if not, feel free to dislike. Stay mysterious until the next video. Take care and bye.
