NVIDIA’s “Blackwell” era has officially arrived, succeeding the RTX 4000 “Ada” generation that has been around since 2022. For some, it’s a bittersweet goodbye; for others, including the owners of RTX 4060 GPUs, it’s a welcome change.
NVIDIA’s new RTX 5000 graphics cards were revealed at CES 2025 along with some impressive claims: can an RTX 5070 outperform an RTX 4090?– and less impressive prices. Let’s take a look at the data NVIDIA shared in its keynote and take a look at when you can expect those GPUs to arrive and where you can buy them on launch day.
January 9, 2025: I updated this rep with all the new information about the RTX 5000 revealed at CES 2025. – Cale Hunt
NVIDIA CEO Jensen Huang officially kicked off CES 2025 with a keynote speech on January 6, where he revealed the company’s next generation of RTX 5000 “Blackwell” graphics cards.
The flagship RTX 5090 and its less rugged brother, the RTX 5080, are expected to launch for the first time on January 30, 2025, followed by the RTX 5070 Ti and RTX 5070 shortly after in February.
By removing the hardware to reveal the Blackwell GPU, NVIDIA lists 92 billion transistors, 4000 AI TOPS, 380 ray tracing TFLOPs, and 125 shader TFLOPs. These are huge numbers. Compared to the latest generation Ada GPUs, the AI TOPS have tripled, while the RT FLOPS have doubled.
Huang made an ambitious claim, saying that the RTX 5070 can match the functionality of the RTX 4090, all for less than a fraction of the price. However, he went on to say that all of this was highly unlikely without AI, which is rarely too surprising.
The new DLSS four and Reflex 2 with Frame Warp proved to be the main drivers of the vast improvement in functionality. Together, they take the functions of those GPUs far beyond what physical hardware can do.
Is that a good thing? It depends on how you look at it. There are concerns floating around about visual fidelity and latency suffering with such a heavy reliance on frame generation, but I’m ultimately reserving judgment until we get the new cards in for testing.
NVIDIA announced the RTX 5090, RTX 5080, RTX 5070 Ti, and RTX 5070 GPUs for desktops and PCs.
The flagship RTX 5090 card unsurprisingly has the most impressive specs, arriving with 32GB of GDDR7 VRAM with 1.792TB/s memory bandwidth. That’s nearly twice as much bandwidth as the RTX 4090, with more and faster GDDR76 VRAM.
Here’s a look at the 4 new cards announced through NVIDIA.
With all the leaks and rumors leading up to NVIDIA’s official announcement, maximum PC enthusiasts were just waiting to know the worth of the new Blackwell GPUs.
As expected, the RTX 5090 is the most expensive, with an MSRP of $1,999. The RTX 5080 drops to $999, followed by the RTX 5070 Ti at $749.
The RTX 5070, which NVIDIA says can accommodate the functionality of an RTX 4090 in components thanks to all the new advances in artificial intelligence, costs $549.
Keep in mind that these are the prices that NVIDIA sets for its reference cards. Custom RTX 5000 GPUs created through NVIDIA partners will likely be priced differently, some higher and some lower than NVIDIA’s MSRP.
The RTX 5090 and RTX 5080 are expected to launch first on January 30, 2025. That includes NVIDIA’s Founders Edition reference cards and custom cards from partners like ASUS, Colorful, Gainward, GALAX, Gigabyte, INNO3D, KFA2, MSI, Palit, PNY, and ZOTAC.
Laptops with cellular versions of those two GPUs are expected in March 2025. Expect to see models from Acer, ASUS, Dell, Gigabyte, HP, Lenovo, MECHREVO, MSI, and Razer.
The RTX 5070 Ti and its sibling RTX 5070 are expected to launch in February 2025.
NVIDIA’s Deep Learning Super Sampling (DLSS) technology has been around since the launch of the first RTX GPUs in 2018, and as the name suggests, it’s been learning and improving the whole time. DLSS 4 is available to any users of RTX 20-series through 50-series cards, but the available features differ due to hardware performance.
For example, the newer RTX 5000 GPUs are the only ones that get the benefits of the new generation DLSS four Multi Frame. The improved DLSS frame generation will also be available only for RTX 50 and Four0 series GPUs. Older forums have access to enhanced DLSS ray reconstruction, DLSS super-resolution, and deep learning smoothing (DLAA).
So what exactly is the next-generation DLSS Four Multi Frame?This is an enhancement of the DLSS Frame Generation that first accompanied DLSS 3, based on the specialized Blackwell GPU architecture for neural rendering capabilities. Using a new AI model, Multi Frame generation takes the generated single frame and multiplies it.
NVIDIA claims that this produces a “consistent and optimal frame rate” despite the extra frames generated. This new generation is largely based on the fifth-generation Tensor cores discovered in Blackwell GPUs, so this feature is not available on RTX 4000 cards. The tool also reduces the use of VRAM, which is a smart thing to do considering that NVIDIA hasn’t surprised anyone with the amount of VRAM.
NVIDIA used Cyberpunk 2077 running with full ray tracing to highlight the improvements. The game sits around 25-30 FPS with DLSS disabled. With DLSS 4 enabled, the frame rate climbs to 235 and beyond, even with ray tracing enabled.
More of the games and programs that already support NVIDIA frame generation will support DLSS 4 multi-frame generation on launch day, and “there are many more on the way. “
NVIDIA’s Reflex generation arrived in 2020 as a tool to reduce latency. Generally, it works by synchronizing the CPU and GPU, resulting in faster popularity of your writing actions. This is especially important in competitive games, but can provide benefits to all players. NVIDIA claims that more than 90% of gamers use Reflex (when available).
New for 2025 is NVIDIA Reflex 2. Another is needed to reduce latency, which NVIDIA calls Frame Warp. Here’s an explanation of how it works provided by NVIDIA itself:
“Reflex 2 Frame Warp takes this concept from research to reality. As a frame is being rendered by the GPU, the CPU calculates the camera position of the next frame in the pipeline, based on the latest mouse or controller input. Frame Warp samples the new camera position from the CPU, and warps the frame just rendered by the GPU to this newer camera position. The warp is conducted as late as possible, just before the rendered frame is sent to the display, ensuring the most recent mouse input is reflected on screen.”
Therefore, moving and distorting a symbol produces missing pixels, which are filled in by a “latency-optimized predictive finishing alpasrhythm. “A lot of technical stuff is starting to go through my head, but at the end of the day?You’ll expect it to be less difficult to aim and track enemies in competitive mode.
Cale Hunt brings over 8 years of delight writing about laptops, PCs, accessories, gaming, and more to Windows Central. If you run Windows or complement hardware, chances are you know about it, have written about it, or are already running it in testing.
Windows Central is from Future US Inc, a global media organization and leading virtual publisher. Visit our corporate site.