Abstract
- RTX 50 graphics playing cards supply a disappointing gaming efficiency uplift over their RTX 40 counterparts whereas being arduous to seek out in inventory and costing a whole lot of {dollars} over MSRP.
- Mid-range and price range RTX 50 GPUs characteristic 12GB and 8GB reminiscence buffers, respectively, which is unacceptable in 2025.
- Some RTX 5090, RTX 5080, and RTX 5070 Ti playing cards are faulty, whereas RTX 5090 and RTX 5080 playing cards additionally undergo from energy connector points that can lead to melted energy cables.
NVIDIA’s newest GPU structure, Blackwell, is spectacular on the subject of the corporate’s information middle product stack. However over on the gaming aspect of the road, issues aren’t as rosy. In truth, RTX 50 GPUs are so underwhelming that I think about RTX 50 probably the most disappointing NVIDIA GPU technology ever.
Grossly Disappointing Efficiency In comparison with the 40 Sequence
NVIDIA has all the time delivered the products relating to efficiency uplift over the earlier GPU technology. For many years, every new GTX, and later RTX, graphics card lineup would blow the earlier technology out of the water.
We had some correct duds that hardly moved the gaming efficiency needle, just like the GeForce 9000 collection. However that was a easy refresh of the Tesla structure that debuted on the GTX 8000 collection. Other than the GTX 9800 GTX and the 9800 GTX+, GeForce 9000 GPUs had been just about the identical by way of gaming efficiency in comparison with their GTX 8000 counterparts.
Even the universally underwhelming generations, just like the RTX 20, delivered the place it mattered. With the RTX 20, we acquired DLSS, and whereas the RTX 2070 did not surpass the earlier technology’s halo product, the legendary GTX 1080 Ti, the RTX 2080 simply hopped over the earlier flagship in gaming efficiency.
Then got here the RTX 40 collection, and one thing was up. Whereas the RTX 4090 was a beast, different members of the family provided considerably disappointing efficiency uplifts harking back to the RTX 20 collection whereas rocking large value bumps in comparison with their RTX 30 equivalents.
If the RTX 40 collection was a disappointment, the RTX 50 lineup is a flat-out catastrophe relating to efficiency enhancements over the earlier technology. The one brilliant spot is the RTX 5090, with the remainder of the stack providing barely any uplift over their predecessors.
The RTX 5080 is barely about 12-15% sooner than the RTX 4080 and miles behind the RTX 4090, one thing remarkable prior to now when the XX80-series GPU would all the time beat the earlier technology flagship. The RTX 5070 Ti surpasses the RTX 4070 Ti by lower than 20%.
The RTX 5070, which does include a fairly strong 23-25% efficiency increase over its predecessor—however is barely about 6% sooner than the RTX 4070 Tremendous—remains to be seen as a disappointment because of NVIDIA’s deceptive advertising and marketing claiming it is as quick because the RTX 5090, the identical 12GB reminiscence buffer discovered the RTX 4070, and poor availability, which leads me to my subsequent level.
Availability Issues Persist
Not having the ability to discover new gaming GPUs at launch is totally regular. Even a decade in the past, you’d have a tough time choosing up a GPU just a few weeks after it debuted. However a few month and a half after it landed, the RTX 5090 remains to be in dismally quick provide. When you in some way discover it, be able to spend the equivalent of an honest used car.
The RTX 5080, which launched the identical day because the RTX 5090, is intermittently in inventory, each on Newegg and Finest Purchase. However you’ll be able to solely get costly OC fashions, which command costs of at the very least a few hundred {dollars} over the cardboard’s $1,000 MSRP.
The RTX 5070 Ti launched only a week after the 2 flagship choices but it surely, too, is uncommon to seek out within the wild. And when you do handle to identify one, you will, once more, need to pay a chunky premium over its $750 MSRP. Lastly, the RTX 5070 was solely launched just a few weeks in the past. So, whereas it is regular to not discover it in inventory, I reckon it, too, will proceed to be arduous to seek out within the upcoming weeks.
Gross sales information confirms that RTX 50 playing cards are arduous to get by. In line with NVIDIA, via The Verge, the RTX 50 collection shipped twice as many items in comparison with the RTX 40 collection throughout the first 5 weeks of availability, however there is a catch.
Whereas we noticed launches of 4 totally different RTX 50 SKUs during the last 5 weeks, the one RTX 40 GPU obtainable throughout the identical interval again in 2022 was the halo mannequin, the RTX 4090.
In a nutshell, NVIDIA claims that it took 4 RTX 50 fashions to double the transport numbers of the RTX 4090, which did not actually fly off the cabinets on account of its $1,600 MSRP. Not less than you would get the RTX 4090 at or close to MSRP for some time. Good luck managing to snag any of the RTX 50 playing cards at MSRP.
Bogus MSRP
It seems like GPU MSRP is worthless today. Not one of the MSRP RTX 50 fashions are presently in inventory, and it seems like they will not reappear on cabinets sooner or later.
When you go to ASUS’s online store, you will notice that you just can not even discover an RTX 5080 mannequin at its $1,000 MSRP. Probably the most reasonably priced possibility, the RTX 5080 Prime, Prime being ASUS’s “MSRP” collection of playing cards, prices exactly $1,264.99 on the time of writing. MSI does not even promote its MSRP Shadow RTX 5080 mannequin anymore. The one Shadow mannequin you’ll find is the pricier OC mannequin that retails for $1,249.99.
The state of affairs with RTX 5090 playing cards is even worse, with probably the most reasonably priced fashions costing a whole lot of {dollars} over the cardboard’s $2,000 asking value (if you’ll find one).
The RTX 5070 Ti and RTX 5070 can sometimes be discovered at MSRP when you’re fast sufficient to order one earlier than bots and scalpers get them, or when you’re fortunate to have a Micro Middle close to you. However contemplating what occurred to the MSRP RTX 5090 and RTX 5080 playing cards, I would not be shocked if we quickly see MSRP RTX 5070 Ti and RTX 5070 listings disappearing for good.
Lastly, it could be finest if I did not even discuss in regards to the state of affairs in Europe, Australia, and different markets, the place MSRPs are even increased, and the MSRP playing cards are nowhere to be discovered as nicely.
Funds and Mid-Vary GPUs Caught With 8GB and 12GB of Reminiscence
Whereas NVIDIA did equip the RTX 5070 Ti with 16GB of reminiscence, the RTX 5070 is caught with a 12GB reminiscence buffer that slowly turns into too little for gaming at 1440p and excessive particulars. Activate ray tracing in trendy AAA video games, and you will practically max out the reminiscence utilization. It seems just like the RTX 5070 can have an analogous destiny to the RTX 3070, whose 8GB of reminiscence was too little for gaming with excessive or extremely visuals shortly after the cardboard got here out.
The state of affairs with the RTX 5060 seems much more dire. Kopite7kimi, a well known leaker with a near-flawless observe document, recently shared the RTX 5060 specs. It seems like the cardboard can have solely 8GB of reminiscence, which is unacceptable in 2025 except we’re speaking a few sub-$200 GPU, which the RTX 5060 definitely will not be. Bear in mind, 8GB of reminiscence debuted on the GTX 1070, which launched virtually 9 years in the past!
My first-hand expertise is that 8GB of reminiscence is not sufficient even for 1080p gaming with max texture high quality, even with out ray tracing. I needed to drop textures to medium in Horizon Forbidden West when streaming the sport to my ASUS ROG Ally to keep away from the large efficiency drop as soon as the reminiscence buffer fills up after about 20 minutes of taking part in the sport with textures set to excessive. I’ve had related experiences in Stalker 2 and Forza Horizon 5, so if RTX 5060 debuts with 8GB of reminiscence, will probably be DOA.
However, an unconfirmed retail listing leak claims the cardboard will include 12GB of reminiscence. I hope that Kopite7kimi is flawed and that the RTX 5060 will certainly pack 12GB of reminiscence. In any other case, anybody who finally ends up shopping for will probably be in for a nasty shock.
Lacking Render Output Models and Melting Energy Cables
The final two nails within the RTX 50 coffin are the lacking ROPs (render output items) and the melting energy cable debacle.
Some RTX 5090, RTX 5080, and RTX 5070 Ti playing cards are defective in that they lack a sure variety of ROPs, making the affected playing cards as much as 10% slower in video games, according to Players Nexus.
This can be a grave oversight by NVIDIA; the corporate in some way failed to note the problem throughout the QA course of. The variety of affected playing cards is small, however that does not cut back the gravity of this fiasco. So as to add insult to damage, whereas NVIDIA claims no cellular RTX 50 GPUs are affected, JarrodTech’s (probably the greatest laptop computer reviewers on the market) sources claim that NVIDIA requested them to double-check their laptops.
Fortunately, it seems just like the RTX 50 cellular GPUs don’t undergo from lacking ROPs. Nonetheless, the truth that NVIDIA requested producers to double-check the laptops does not paint a fairly image relating to its QA course of.
The second flop consists of melting power cables on some RTX 5090 and RTX 5080 graphics playing cards. Now, the 12VHPWR energy connector NVIDIA used on RTX 40 GPUs was already pretty darn bad, however the 12V-2×6 connector that changed it is even worse.
Energy cables will proceed to soften except NVIDIA replaces the 12V-2×6 connector or redesigns RTX 50 circuit boards to incorporate a greater fuse safety answer. For the time being, every part factors out that NVIDIA will do neither.
The only real silver lining could be that RTX 5090 playing cards have been nowhere to be discovered for weeks now. There’s an opportunity the corporate is attempting to unravel the problem below the radar, and that the long run RTX 5090 playing cards will not include shoddy energy connectors.
As for myself, I’ve gotten an AMD RX 9070 XT instead of an RTX 5070 Ti. The AMD card gives extra worth together with a wonderful upscaling answer within the type of FSR 4, which lastly matches DLSS in image high quality.
If NVIDIA delivers extra bang for the buck in comparison with AMD with its subsequent technology of gaming GPUs (I significantly doubt this), I am going to change again as a result of I maintain no love for any model. However the reality of the matter is that the RTX 50 collection is an enormous disappointment, the largest from NVIDIA so far, so far as I am involved.