Nvidia has announced its new line of GPUs, the RTX 50-series. That includes four new graphics cards: the RTX 5090, RTX 5080, RTX 5070 Ti, and RTX 5070. They’re nearly as insane in price and power as the rumors suggested. Nvidia is promising some huge leaps in performance, thanks to the new AI powers of DLSS 4.
Here’s everything we know so far.
RTX 50-series: pricing and release date
To say there was speculation about the price of the RTX 50-series is a serious understatement. Well, the leaks and rumors can finally come to an end. Here are the prices of upcoming GPUs:
- RTX 5090: $1,999
- RTX 5080: $999
- RTX 5070 Ti: $749
- RTX 5070: $549
Get your weekly teardown of the tech behind PC gaming
That’s a $400 increase on the top end, up from the $1,599 for the RTX 4090, which was already the most expensive Nvidia GPU ever made. That price jump is definitely an eyebrow-raiser, but further down the stack, the pricing actually isn’t bad.
The RTX 5080 matches the price of the RTX 4080 (or RTX 4080 Super), for example. Meanwhile, the RTX 5070 Ti and RTX 5070 are actually $50 cheaper than the RTX 4070 Ti and RTX 4070. How’s that for a surprise?
The RTX 5090 and 5080 both launch on January 30. The RTX 5070 Ti and 5070, meanwhile, will follow sometime in February.
RTX 50-series: specs
| RTX 5090 | RTX 5080 | RTX 5070 Ti | RTX 5070 | |
| Architecture | Blackwell GB202 | Blackwell GB203 | Blackwell GB203 | Blackwell GB205 |
| CUDA cores | 21760 | 10752 | 8960 | 6140 |
| Tensor cores | 5th-gen 3352 TOPS | 5th-gen 1801 TOPS | 5th-gen 1406 TOPS | 5th-gen 988 TOPS |
| Ray tracing cores | 4th-gen 318 TFLOPS | 4th-gen 171 TFLOPS | 4th-gen 133 TFLOPS | 4th-gen 94 TFLOPS |
| Boost clock (GHz) | 2.41 | 2.62 | 2.45 | 2.51 |
| Base clock (GHz) | 2.01 | 2.30 | 2.30 | 2.16 |
| Max resolution | 4K at 480Hz or 8K at 120Hz | 4K at 480Hz or 8K at 120Hz | 4K at 480Hz or 8K at 120Hz | 4K at 480Hz or 8K at 120Hz |
| Memory interface width | 512-bit | 256-bit | 256-bit | 192-bit |
| Memory | 32GB GDDR7 | 16GB GDDR7 | 16GB GDDR7 | 12GB GDDR7 |
| Memory speed | 28Gbps | 30Gbps | 28Gbps | 28Gbps |
| TOPS | 3,400 | 1,800 | 1,400 | 1,000 |
| TGP (watts) | 575 | 360 | 300 | 250 |
| Required system power (watts) | 1,000 | 850 | 750 | 650 |
| Price | $1,999 | $999 | $749 | $549 |
There’s a lot to take in here. We also know that the cards will all use a 16-pin connector and DisplayPort 2.1, which is where the support for 8K resolutions at 120Hz comes from. The cards are also using the 12V-2×6 power connector. This can almost be considered overkill for most of the lineup, but the 575-watt RTX 5090 may not have a lot of overclocking headroom with a single 600-watt power connector.
While we have the official specs, there’s a few more details that we likely won’t have confirmed until reviewers get their hands on the cards.
The initial announcement didn’t deliver some key information, including memory speed and process node, which Nvidia has since supplied. As a result, we know that the RTX 5080 will have faster memory modules than any of the other GPUs. This has long been a subject of speculation, and has now been confirmed by Nvidia.
Overall, we’re looking at gains all across the board here in many specs, but not in all of them. Unsurprisingly, the RTX 5090 scores the biggest wins, with a huge memory increase (to a whopping 32GB, no less) and a boost to CUDA cores. The other cards can’t enjoy the same kind of treatment, with all three retaining the same VRAM and bus configuration. However, the entirety of the RTX 50-series uses much faster GDDR7 memory, which is an upgrade in itself.
How these specs translate to real-world performance will be revealed only once we get to review these GPUs ourselves.
RTX 50-series: mobile GPUs
Somewhat unexpectedly, Nvidia’s keynote wasn’t solely focused on desktop GPUs. In fact, gaming laptops played a big part in the presentation, and there’s a lot for notebook gamers to get excited about.
Nvidia and its partners are readying up laptops with almost the full range of RTX 50-series cards, including the RTX 5090, RTX 5080, RTX 5070 Ti, and the RTX 5070. The RTX 5060 Ti, RTX 5060, and the RTX 5050 haven’t been mentioned yet, but in all likelihood, some (if not all) will be coming at a later date.
The only thing we know about these new laptop GPUs is the AI tera operations per second (TOPS) count for each, starting with 1,850 TOPS for the RTX 5090 (which is comparable to the desktop RTX 5080) and ending with 800 AI TOPS for the RTX 5070. That laptop is meant to be mighty impressive, though, as Huang claims that it’ll rival the performance of an RTX 4090. Desktop or laptop version? It’s unclear right now, but I’m guessing laptop.
Nvidia has also shared the starting prices for each laptop. The RTX 5070 models start at $1,299, followed by $1,599 and up for the RTX 5070 Ti and $2,199 for the RTX 5080. Laptops equipped with the RTX 5090 get really expensive, with the cheapest models priced at $2,899.
Many of these laptops will become available in March. Nvidia’s partners, such as MSI, have already teased some of their upcoming offerings.
RTX 50-series: design
| RTX 5090 | RTX 5080 | RTX 5070 Ti | RTX 5070 | |
| Length | 304mm | 304mm | Varies | 242mm |
| Width | 137mm | 137mm | Varies | 112mm |
| Slot | 2-slot | 2-slot | Varies | 2-slot |
| Power connector | 4x PCIe 8-pin cables or 1x 600 W PCIe Gen 5 cable | 3x PCIe 8-pin cables or 1x 450 W or greater PCIe Gen 5 cable | 2x PCIe 8-pin cables or 300 W or greater PCIe Gen 5 cable | 2x PCIe 8-pin cables or 300 W or greater PCIe Gen 5 cable |
The Founders Edition RTX 5090 might be the biggest surprise with its 2-slot design. In fact, the RTX 5090 and 5080 are the exact same size, much like the bigger RTX 4080 and 4090. The new design still uses two fans, but they’re now on the same side, unlike the previous generation, which used a dual axial flow design.
Like in prior years, there is no Founders Edition of the RTX 5070 Ti.
Of course, board partners will no doubt make much larger versions of these cards that take up more than two slots. In fact, so far, the Founders Edition of the RTX 5090 is the only one that’s ready for Nvidia’s SFF program, implying that partner cards are going to be significantly larger.
RTX 50-series: architecture
We know for a fact that the follow-up to Ada Lovelace is called Blackwell, honoring American mathematician David Blackwell. The new consumer GPUs are manufactured based on TSMC’s 4NP process, which is a 4nm node. Early leaks about the cards suggested that Nvidia would switch to a 3nm process, but that didn’t pan out.
The lineup includes chips starting from the high-end GB202 through the GB203, GB205, GB206, and entry-level GB207. We haven’t seen the GB207 yet, though. The RTX 5070 runs on the GB205 architecture, so presumably, the RTX 5060 will use GB207, but Nvidia has yet to announce the budget part of the lineup.
The exact architecture of the Blackwell consumer GPUs is still a bit of an unknown. Although Nvidia shared the specs for each card and detailed what makes them better than the previous generation, not everything is crystal clear yet — but it won’t stay that way much longer. With the first two GPUs less than a month away, we’ll learn more as the release date draws closer.
However, Nvidia divulged some more information during CES 2025. As shared by Tom’s Hardware, the tech giant had to completely redefine the printed circuit board (PCB) to support the kind of power consumption (and just overall power, really) that the new architecture can provide. The GPU that Nvidia talked about was, of course, the RTX 5090 with the GB202 chip.
The PCB is split into three parts, with the main PCB equipped with the massive 30-phase VRM, the copious amounts of VRAM, and the GB202 chip. This is the only part of the PCB that Nvidia unveiled during the presentation, which — although small — still houses a powerful 92-billion-transistor GPU. The missing pieces include display outputs, the PCIe grid, and power connectors for fans. It’s truly a design we have never seen before.
Considering that even the monstrous RTX 5090 is just a dual-slot card, Nvidia had to do more than just implement a new PCB design. It also gave the cooling system a bit of an overhaul, using a liquid metal thermal interface material (TIM) instead of thermal paste.
Nvidia also introduced RTX Neural Shaders that bring AI to programmable shaders. This, in turn, will boost lighting and other quality-related features in games, all in real time.
Although Blackwell is still so new, it’s already seen its fair share of troubles. A recent report from The Information tells us that Nvidia’s data center GPUs are running into problems with overheating, which has resulted in delays in the deployment of server racks used for AI training. Let’s hope that with the wildly different form factor and the considerate approach to cooling, these problems will never make it to consumer GPUs.
RTX 50-series: DLSS 4
Jensen Huang says that Nvidia’s Deep Learning Super Sampling 4 can “see into the future.” It’s a bold claim, but based on early information, the updated upscaling and frame generation tech sounds like it’ll play a large role in the success of this new generation of GPUs.
DLSS 4, and its neural rendering capabilities, is how Huang can explain the fact that the $550 RTX 5070 can reportedly rival the $1,600 RTX 4090. During the keynote, Huang explained that most of the pixels are generated using Nvidia’s Tensor cores, and as a result, only the required pixels are rendered, while the majority is generated with AI.
“The future of computer graphics is neural rendering,” Huang said.
DLSS 4 takes the frame generation we’ve grown to know (and mostly enjoy) with DLSS 3 and kicks it up a couple of notches. The latest update introduces Multi Frame Generation, which, as Nvidia says, will boost gaming performance by generating up to three frames for each rendered frame.
This is said to boost frames per second (fps) by up to eight times, and Nvidia’s example benchmarks show that. The RTX 5090 is said to be able to hit an impressive 240 fps at 4K with full ray tracing, which presumably also includes path tracing.
Nvidia first introduced frame generation with DLSS 3, but DLSS 4 comes with a new, significantly faster AI model that cuts back on the computational cost to deliver better performance. The model is said to be up to 40% faster while also being up to 30% less VRAM-intensive. Nvidia also got rid of the hardware-based optical flow we’ve seen in the RTX 40-series and replaced it with an AI model.
DLSS 4 is still a taxing thing for the GPU to run, but Nvidia is tackling the challenge with hardware improvements, too. It introduced fifth-gen Tensor cores, which are said to deliver up to 2.5x the AI processing power of the previous generation. There’s also hardware Flip Metering and an improved display engine.
Overall, DLSS 4 will generate 15 out of every 16 pixels.
Set to come out with the new GPUs at the end of January, DLSS 4 will see decent adoption from the get-go. Around 75 games will support DLSS 4, and this includes some of the AAA titles that will benefit from it the most, such as Indiana Jones and the Great Circle and Cyberpunk 2077.
RTX 50-series: performance
Nvidia made some significant claims about the performance of these new 50-series cards, saying they double the performance of their predecessors. As you can see in the slides above, that includes games, yes, but also AI and creative applications. The only direct comparison Nvidia made was for the RTX 5070, which it claims is now on par with the RTX 4090.
That sounds like quite a leap, but remember: it’s done on the back of DLSS 4’s multi-frame generation feature. Most of Nvidia’s performance claims and graphs include DLSS 4, so it’s difficult to know exactly what the performance delta is when it comes to pure rendering.