The Nvidia GeForce RTX 50 graphics cards are available, starting with the RTX 5090 and RTX 5080. Price, release date, technical specifications, performance, Blackwell architecture, DLSS, ray tracing, new features and improvements, here is everything we know about the latest generation of Nvidia GPUs.
The long-awaited Nvidia GeForce RTX 50 series of graphics cards is finally here. These new GPUs are based on the new Blackwell architecture, which replaces the previous generation's Ada Lovelace. With this renewal of the range, the increase in power and performance is coupled with better energy efficiency. But all this comes at a high price.
When will RTX 50 graphics cards be available?
The RTX 5080 and RTX 5090 were revealed at CES 2025 in Las Vegas. Nvidia always launches its highest-end SKUs first, so the RTX 5090 and RTX 5080 arrived first on January 30, 2025.
Nvidia unveiled the RTX 5070 and 5070 Ti in early February. The RTX 5070 Ti was expected to be released on February 20, 2025, followed by the standard version in March. But according to a reliable leaker, Nvidia has pushed back the arrival of the two models to March 2025. A decision possibly taken to compete directly with AMD's Radeon RX 9070, scheduled for the same month, or because of difficulties in the production of the cards.
The RTX 5060 and 5060 Ti could arrive in April 2025, but we'll have to wait to be sure. As for a desktop RTX 5050, we know that it exists. It will be the last to arrive in stores, in May 2025 at the earliest. A laptop version is also expected to see the light of day.
How much do RTX 50 graphics cards cost?
Until Nvidia has announced the price of all its RTX 50s, we can only rely on guesses and leaks. And yet, even when Nvidia announces the prices, we know that Founders Edition graphics cards are rare, and that you often have to turn to models customized by third-party manufacturers like Gigabyte, Asus, PNY or MSI to equip yourself, who often charge higher prices than those of Nvidia.
For the moment, we know the following prices:
- RTX 5090 Founder's Edition: €2,349.
- RTX 5080 Founder's Edition: €1,179.
On the Nvidia website, the RTX 5070 Ti is announced starting at €884. A strange price that seems far from reflecting reality since French resellers list it between €1,149 and €1,199. We have to wait for confirmation, but already the hopes of a card for less than €1,000 are starting to fly away.
What is Blackwell?
The Blackwell architecture was announced to Nvidia investors in late 2023 and was made official to the general public in March 2024. It was presented as the company's new GPU architecture for data centers (in particular for servers dedicated to machine learning and generative AI computing, replacing the Hopper architecture. Surprise, Blackwell is also the architecture used for consumer graphics cards, thus succeeding Ada Lovelace. On the other hand, we distinguish the GB100 chips for data centers from the GB200 chips for RTX 50 graphics cards.
The Blackwell architecture, and therefore the GB200 GPU chips of the RTX 50, is based on TSMC's 4NP manufacturing process. We are moving to a 4 nm node on this generation (compared to 5 nm on the RTX 40), while TSMC has already developed a 3nm engraving. Nvidia preferred to start with an improved version of an already well-mastered process rather than starting with a completely new technology. It is not known whether this choice was made for reasons of performance, reliability, costs, production capacity, or a combination of several of these factors.
What are the performances of the RTX 50?
We were able to put two graphics cards from the RTX 50 series to the test. We invite you to read our tests to get a precise idea of the capabilities of the components:
- Test of the RTX 5090
- Test of the RTX 5080
- Test of the RTX 5070 Ti
While waiting for those of the RTX 5070 and 5070 Ti, we can already say that according to Nvidia, the RTX 5070 Ti offers a memory bandwidth of 896 GB/s, an increase of 78% compared to the RTX 4070 Ti. The standard version, despite a reduced bus, sees its bandwidth increase by 33% thanks to GDDR7. In terms of frequencies, the RTX 5070 reaches 2.51 GHz, slightly more than the RTX 4070, while the Ti displays 2.45 GHz, down from its predecessor.
What are the technical characteristics of the RTX 50?
One of the major new features brought by the RTX 50 is the move to GDDR7 memory, while the RTX 40 remained in GDDR6 or GDDR6X. Pushed to its limits, GDDR7 can theoretically reach speeds of 36 Gb/s, but Nvidia doesn't go that far.
The RTX 5090 features 32 GB of GDDR7 VRAM for a 512-bit bus, which translates into an exceptional bandwidth of approximately 1.8 TB/s. Clearly, this is a product that is no longer intended for video games, but for professional uses requiring more resources.
The RTX 5080 benefits from 16 GB of GDDR7 VRAM and a 256-bit bus. It is mainly intended for gamers aiming for 4K and ray-tracing without any compromise. Be careful though: some models are less efficient than others and it is very difficult to realize this without comparing them with others.
The RTX 5070 Ti shares the same GB203 GPU as the RTX 5080 and integrates 16 GB of GDDR7 memory. The RTX 5070, on the other hand, is based on the GB205 GPU, the least powerful in the RTX 5000 range, but has a higher number of CUDA, Tensor and RT cores than the 4070. It has 12 GB of GDDR7 with a 192-bit bus.
Quelle connectorique pour les RTX 50 ?
Goodbye PCIe 4, the RTX 50 are compatible with the latest PCIe 5 hardware interface standard, with a 16-pin connector. Don't panic if your motherboard is not very recent, backward compatibility with the PCIe 4 standard is well ensured.
Everything has been thought of so that the RTX 5090 does not catch fire like the 4090, but unfortunately, connectors on the power cable have still melted. Some have made warning systems to be able to turn off the PC in the event of overheating, but we would prefer that Nvidia solve the problem.
To connect display devices to your machine, the RTX 50s of course offer HDMI 2.1 connectivity. Several DisplayPort ports are also available, with a switch to DisplayPort 2.1 on this generation. The RTX 40s are stuck at DisplayPort 1.4, while AMD's Radeon RX 7000s have been given the famous DisplayPort 2.1. The latter grants the possibility of transmitting a 4K UHD stream at 480 Hz and 8K at 165 Hz with DP80 certified cables, which support UHBR20 (Ultra-High Bit Rate) at 80 Gb/s bandwidth.
0 Comments