Nvidia GeForce RTX 5090 discontinued to be replaced by RTX 5080 24GB, says rumor

So long RTX 5090, hello RTX 5080 24GB

The Nvidia GeForce RTX 5090 is no more. Yes, it finally happened. Nvidia decided it simply wasn’t making enough money from selling flagship GPUs to gamers and has instead decided to funnel all of its top chip production purely into making AI chips instead. Or at least, that’s the picture painted by the latest rumors.

According to regularly reliable tech leaker MegasizeGPU, all of Nvidia’s GB202 GPU production has now been funnelled away from making RTX 5090s and instead all these chips will be used to produce RTX Pro 6000 cards for AI workloads. In the 5090’s place, gamers will apparently be getting an uprated RTX 5080 with 24GB of VRAM, up from 16GB in the current model.

If true, this really is wild. It’s one thing for RTX 5090 prices to be going absolutely crazy right now due to high demand and scarcity, but it’s another for Nvidia to pull the plug entirely.

For context, RTX 5090 prices have reached record highs in recent weeks, with them being sold for double, triple or even quadruple their original $2000 MSRP. As such, you might be sat there thinking that surely Nvidia is making huge amounts of money even on its RTX 5090 cards. However, while it’s certainly making a profit, the reality is that its RTX Pro 6000 cards still sell for far, far more.

A quick look on Newegg shows us that RTX 5090s are selling for as much as $9000, but that’s still well short of the $16000 that Nvidia charges for an RTX Pro 6000. It’s also worth noting that the RTX Pro 6000 has itself double in price since its launch.

Now, again, this apparent complete abandonment of the RTX 5090 is all based on rumors, but along with MegasizeGPU having a decent track record for GPU-related leaks, this does all just make sense, unfortunately. In fact, it’s almost surprising it has taken Nvidia this long to make this move.

That’s because the entire chip production industry has been operating at capacity for so long now that every company like Nvidia has to absolutely maximise its profits from the allocation of production it can get. And, when it comes to chips, few are more expensive than the GB202, simply because it’s so huge. Covering an area of approximately 750mm2, it’s nearly double the size of the RTX 5080’s GB203 GPU. In turn, this means Nvidia can only produce roughly 70 or so GB202 dies per 300mm silicon wafer, compared to around 150 GB203 dies. With so few chips to work with, it simply makes sense to maximise your profit on them.

Now, the crucial thing to understand here is that Nvidia isn’t necessarily benefitting as much as you might expect from the current inflated RTX 5090 prices. Those price hikes are supply and demand on the consumer/retailer side, whereas Nvidia will have been charging a set, much lower figure for its chips to its partners that make its graphics cards, even if that figure might have risen somewhat in recent months.

In contrast, the RTX Pro 6000 is a chip that the company will have already been charging much higher rates for.

So, where does this leave gamers and other home users? Well, as mentioned, there is expected to be an RTX 5080 with 24GB of VRAM coming soon, but it’s not known when that will arrive, whether it will also include any clock speed increases for extra performance, and how much Nvidia will charge for it. It’s possible Nvidia will shuffle its whole graphics card stack down a notch, replacing the RTX 5080 at $1000 with the 24GB RTX 5080, but given current graphics card prices across the board, this seems unlikely. Ultimately, we’ll just have to wait and see on that front.

As for future launches, while this could be seen as a sign of Nvidia pushing more and more towards server-powered world where compute power is rented and noone owns a PC that has any processing power of its own, I don’t think we’re quite there yet. After all, the company is currently pushing its neural rendering DLSS 5 technology, and that is so demanding that it effectively is only aimed at high end PC owners. Plus, there will absolutely continue to be certain types of games and other applications where local powerful computing will be required.

Instead, this could all be a sign of the company finally transitioning to its next generation of GPUs. Current rumors put a possible launch of the RTX 6090, or whatever it ends up being called, towards the latter half of next year, so we’re still a fair way off, but if Nvidia is ramping up that production, it would make sense to maximise profits on the last of its GB202 GPUs while it can.

It may also be the case that Nvidia is well aware that most gamers and real home users simply aren’t buying these cards anymore. Their prices are already so ludicrous that it’s effectively madness not only to buy one, but to even keep hold of one if you have one. Here in the UK, I could sell this for £5000, which is roughly $7000, buy an RTX 5080 and still have more three grand left over.

So, while it looks and feels wrong to us, it’s probably fair to say that it will actually have minimal impact on the market as a whole.

That said, I’d be interested to hear from you in the comments. Do you genuinely feel like you would have wanted to buy an RTX 5090 in the near future? Have you already sold your 5090? And are any of you the people that are actually buying these cards at current prices? If so, what you using them for? For now, though, from me and the RTX 5090, it’s time to say goodbye.

Add a Comment

Your email address will not be published. Required fields are marked *