GeForce 50 series
| Release date | January 30, 2025 |
|---|---|
| Manufactured by | TSMC |
| Designed by | Nvidia |
| Marketed by | Nvidia |
| Architecture | Blackwell |
| Fabrication process | TSMC 4NP |
| Cards | |
| Mid-range |
|
| High-end |
|
| Enthusiast |
|
| History | |
| Predecessor | GeForce 40 series |
| Support status | |
| Supported | |
The GeForce 50 series is an upcoming series of consumer graphics processing units developed by Nvidia, succeeding the GeForce 40 series. Announced at CES 2025, it will debut with the release of the RTX 5080 and RTX 5090 on January 30, 2025.
The cards are based on Nvidia's Blackwell architecture and feature fourth-generation RT cores for hardware-accelerated real-time ray tracing, and fifth-generation deep-learning-focused Tensor Cores.
Background
[edit]In March 2024, Nvidia announced the Blackwell architecture for its datacenter products. Like Ampere, Blackwell is a shared architecture between both consumer and datacenter products rather than distinct architectures released simultaneously like Ada Lovelace for consumers and Hopper for datacenter.
At the Game Awards in December 2024, a cinematic trailer for The Witcher IV was shown which had been pre-rendered on an "unannounced Nvidia GeForce RTX GPU". This was assumed to be an upcoming GeForce 50 series GPU.[1][2] Later in the same month, it was reported that Nvidia had begun stockpiling GeForce 50 series units in U.S. warehouses due to the looming potential of tariffs.[3][4] Donald Trump during his 2024 presidential campaign floated a blanket 10% tariff on all imports and a 60% tariff on imports from China.[5][6]
Announcement
[edit]
On January 6, 2025, the GeForce 50 series was officially announced for both desktop and mobile devices during Nvidia's CES keynote in Las Vegas.[7] The pricing announcement was met with surprise as the RTX 5080 at $999 was the same price that the RTX 4080 Super released at a year earlier despite the anticipated price increases.[8] Nvidia CEO Jensen Huang claimed that the RTX 5070 could reach "RTX 4090 performance at $549" despite a reliance on DLSS 4 upscaling and multi-frame generation rather than a direct comparison of raw compute.[9]
Features
[edit]- Up to 21,760 CUDA cores
- Up to 32 GB of GDDR7 VRAM
- PCIe 5.0 interface
- DisplayPort 2.1a and HDMI 2.1a display connectors
Blackwell architecture
[edit]The GeForce 50 series is powered by the Blackwell microarchitecture which continues Ada Lovelace's emphasis on high graphics frequencies and large L2 caches. The Blackwell architecture introduces Nvidia RTX's fourth-generation RT cores for hardware-accelerated real-time ray tracing and fifth-generation Tensor Cores for AI compute and performing floating-point calculations.
GDDR7
[edit]| Bus width | Theoretical bandwidth (GB/s) | |
|---|---|---|
| GDDR6X (21 Gbps) |
GDDR7 (28 Gbps) | |
| 256-bit | 672 | 896 |
| 384-bit | 1,008 | 1,344 |
| 512-bit | 1,344 | 1,792 |
GeForce 50 series GPUs are the first consumer GPUs to feature GDDR7 video memory for greater memory bandwidth over the same bus width compared to the GDDR6 and GDDR6X memory used in the GeForce 40 series. GeForce 50 series desktop GPUs use GDDR7 modules from Samsung due to them being available for validation earlier than modules from SK Hynix and Micron.[10][11]
12V2×6 connector
[edit]The GeForce 50 series uses the 16-pin 12V2×6 connector which is a revision of the 12VHPWR connector featured on the GeForce 40 series. There were problems with the 12VHPWR connector melting on some RTX 4090 GPUs due to the connector not being fully seated and connector design flaws that did not implement a high enough safety and error tolerance.[12] The 12V2×6 connector revision, published by PCI-SIG in July 2023, addressed this by shortening the four sense pins so the connector will not push any power if it has not been fully seated.[13] The 12VHPWR design would still draw up to 150W of power even if the sense pins were not making full contact. 12V2×6 is backwards compatible with existing 12VHPWR cables and adapters.
Nvidia has mandated to its AIB partners that the 16-pin 12V2×6 connector be used on all GeForce 50 series designs.[14] With the GeForce 40 series, the 12VHPWR connector was only mandated on higher power SKUs such as the RTX 4070 Ti, RTX 4080 and RTX 4090 while RTX 4060, RTX 4060 Ti and RTX 4070 AIB designs had the option of using 8-pin PCIe connectors. The 600W-capable 12VHPWR connector would not have been necessary on sub-200W SKUs.
DLSS 4
[edit]The fourth generation of Deep Learning Super Sampling (DLSS) was unveiled alongside the GeForce 50 series. DLSS 3 upscaling and frame generation was exclusive to GeForce 40 series GPUs. DLSS 4 upscaling uses a new vision transformer-based model for enhanced image quality with reduced ghosting and greater image stability in motion compared to the previous convolutional neural network (CNN) model.[15] DLSS 4 allows a greater number of frames to be generated and interpolated based on a single traditionally rendered frame. This form of frame generation called Multi-Frame Generation is exclusive to the GeForce 50 series while the GeForce 40 series is limited to one interpolated frame per traditionally rendered frame. Nvidia claims that DLSS 4's frame generation model uses 30% less video memory with the example of Warhammer 40,000: Darktide using 400MB less memory at 4K resolution with frame generation enabled.[16] Nvidia claims that 75 games will integrate DLSS 4 Multi-Frame Generation at launch, including Alan Wake 2, Cyberpunk 2077, Indiana Jones and the Great Circle, and Star Wars Outlaws.[17]
| GeForce 20 series | GeForce 30 series | GeForce 40 series | GeForce 50 series | |
|---|---|---|---|---|
| Transformer Model | 
|
|
|
|
| 2x Frame Generation | 
|
|
|
|
| 3-4x Frame Generation |
|
|
|
|
I/O
[edit]The GeForce 50 series includes DisplayPort 2.1a with higher display output data rates to support high resolution and high refresh rate displays. The GeForce 40 series received criticism for only including DisplayPort 1.4 (32Gbps) while the competing Radeon RX 7000 series included DisplayPort 2.1 UHBR13.5 (54Gbps).
At CES 2025, VESA announced a collaboration with Nvidia on the new DP80LL ("low loss") UHBR20 active cable standard.[18] DP80LL allows for 80Gbps DisplayPort 2.1 cables up to 3 meters long as passive DP80 cables are limited in length due to signal integrity concerns.
Products
[edit]Desktop
[edit]GeForce 50 series desktop GPUs are the first consumer GPUs to utilize a PCIe 5.0 interface and GDDR7 video memory.
| SKU | Release date | Launch MSRP (USD) |
GPU die |
Transistors (billion)
|
Die size |
Core | SMs | Cache | Memory | Fillrate[a][b] | Processing power (TFLOPS) | Interface | TDP | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Config[c] | Clock (MHz)[d] |
L1 | L2 | Type | Size | Clock (Gb/s) |
Band- width (GB/s) |
Bus width |
Pixel (Gpx/s) |
Texture (Gtex/s) |
FP16 | FP32 | FP64 | Tensor compute [sparse] | |||||||||
| GeForce RTX 5070 |
Feb 2025 | 549 | GB205-300 | ? | 6,144 192:96:48:192 |
2160 2510 |
48 | 6 MB | MB | GDDR7 | 12 GB | 28.0 | 672 | 192-bit | PCIe 5.0 x16 | 250 W | |||||||
| GeForce RTX 5070 Ti |
749 | GB203-300 | 377 mm2 | 8,960 280:140:70:280 |
2300 2450 |
70 | 8.75 MB | MB | 16 GB | 896 | 256-bit | 300 W | |||||||||||
| GeForce RTX 5080 |
Jan 30, 2025 | 999 | GB203-400 | 10,752 336:168:84:336 |
2300 2620 |
84 | 10 MB | MB | 30.0 | 960 | 360 W | ||||||||||||
| GeForce RTX 5090 |
1,999 | GB202-300 | 92 | 744 mm2 | 21,760 680:340:170:680 |
2010 2410 |
170 | 21.25 MB | MB | 32 GB | 28.0 | 1,792 | 512-bit | 575 W | |||||||||
- ^ Pixel fillrate is calculated as the number of render output units (ROPs) multiplied by the base (or boost) core clock speed.
- ^ Texture fillrate is calculated as the number of texture mapping units (TMUs) multiplied by the base (or boost) core clock speed.
- ^ CUDA Cores
Texture Mapping Units: Render Output Units: Ray Tracing Cores: Tensor Cores - ^ Core boost values (if available) are stated below the base value in italics.
Mobile
[edit]Laptops featuring GeForce 50 series laptop GPUs were shown at CES 2025. Laptops with GeForce 50 series GPUs were paired with Intel Arrow Lake-HX and AMD Strix Point and Fire Range CPUs.
| SKU | Release date | GPU Die |
Transistors (billion)
|
Die size |
Core | SMs | Cache | Memory | Fillrate[a][b] | Processing power (TFLOPS) | Interface | TDP | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Config[c] | Clock (MHz)[d] |
L1 | L2 | Type | Size | Clock (Gb/s) |
Band- width (GB/s) |
Bus width |
Pixel (Gpx/s) |
Texture (Gtex/s) |
FP16 | FP32 | FP64 | Tensor compute [sparse] | ||||||||
| GeForce RTX 5070 Laptop |
Mar 2025 | GB206-300 | ? | 4,608 144:72:36:144 |
36 | 4.5 MB | MB | GDDR7 | 8 GB | 28.0 | 448 | 128-bit | PCIe 5.0 x16 | 50-100 W | ||||||||
| GeForce RTX 5070 Ti Laptop |
GB205-300 | ? | 5,888 184:92:46:184 |
46 | 5.75 MB | MB | 12 GB | 672 | 192-bit | 60-115 W | ||||||||||||
| GeForce RTX 5080 Laptop |
GB203-300 | 377 mm2 | 7,680 240:120:60:240 |
60 | 7.5 MB | MB | 16 GB | 896 | 256-bit | 80-150 W | ||||||||||||
| GeForce RTX 5090 Laptop |
GB202-300 | 92 | 744 mm2 | 10,496 336:128:84:336 |
82 | 10.25 MB | MB | 24 GB | 1,344 | 384-bit | 95-150 W | |||||||||||
- ^ Pixel fillrate is calculated as the number of render output units (ROPs) multiplied by the base (or boost) core clock speed..
- ^ Texture fillrate is calculated as the number of texture mapping units (TMUs) multiplied by the base (or boost) core clock speed.
- ^ CUDA Cores
Texture Mapping Units : Render Output Units : Ray Tracing Cores : Tensor Cores - ^ Core boost values (if available) are stated below the base value in italics.
References
[edit]- ^ Chacos, Brad (December 13, 2024). "Nvidia stokes RTX 50-series hype with Witcher 4 and a global LAN party". PCWorld. Retrieved January 6, 2025.
- ^ Nasir, Hassan (December 13, 2024). "Nvidia teases RTX 50 Blackwell Gaming GPUs for launch next month — The Witcher IV's first cinematic trailer likely leveraged the upcoming RTX 5090". Tom's Hardware. Retrieved January 6, 2025.
- ^ Zuhair, Muhammad (December 26, 2024). "Nvidia & AMD Rush To Ship Out Next-Gen GPUs To Avoid Trump Tariffs; GeForce RTX 5090 Estimated To Cost $2,500+". Wccftech. Retrieved January 6, 2025.
- ^ Garreffa, Anthony (December 26, 2024). "AMD and Nvidia ship next-gen gaming GPUs from China to US before Jan 20 to escape tariffs". TweakTown. Retrieved January 6, 2025.
- ^ Arantani, Lauren (October 15, 2024). "Trump vows to impose tariffs as experts warn of price hikes and angry allies". The Guardian. Retrieved January 6, 2025.
Trump is proposing an at least 10% blanket tariff on all imports, with tariffs as high as 60% on goods from China.
- ^ Chu, Ben (October 15, 2024). "Would Donald Trump's tariffs hurt US consumers?". BBC News. Retrieved January 6, 2025.
- ^ Warren, Tom (January 6, 2025). "Nvidia announces next-gen RTX 5090 and RTX 5080 GPUs". The Verge. Retrieved January 6, 2025.
- ^ Chacos, Brad (January 6, 2025). "Surprise! Nvidia's GeForce RTX 50-series GPUs cost less than you thought". PCWorld. Retrieved January 7, 2025.
- ^ James, Dave (January 6, 2025). "Nvidia's new RTX 5070 will deliver 'RTX 4090 performance at $549' when it launches in February". PC Gamer. Retrieved January 7, 2025.
- ^ "Nvidia reportedly prioritizes Samsung GDDR7 memory for desktop GeForce RTX 50 series". VideoCardz. November 25, 2024. Retrieved January 6, 2025.
- ^ Mujtaba, Hassan (November 25, 2024). "Nvidia GeForce RTX 50 "Blackwell" GPUs Reportedly Utilize Samsung's GDDR7 Memory Chips". Wccftech. Retrieved January 6, 2025.
- ^ Mah Ung, Gordon (June 15, 2023). "Melting GeForce RTX 4090 power cables: A timeline of events". PCWorld. Retrieved January 6, 2025.
- ^ Liu, Zhiye (July 3, 2023). "16-Pin Power Connector Gets A Much-Needed Revision, Meet The New 12V-2x6 Connector". Tom's Hardware. Retrieved January 6, 2025.
- ^ Campbell, Mark (February 19, 2024). "Nvidia to mandate 16-pin PCIe 6.0 power with RTX 50 series "Blackwell" GPUs". OC3D. Retrieved January 6, 2025.
- ^ Leadbetter, Richard (January 7, 2025). "Hands-on with DLSS 4 on Nvidia's new GeForce RTX 5080". Eurogamer. Retrieved January 7, 2025.
- ^ Lin, Henry; Burnes, Andrew (January 6, 2025). "Nvidia DLSS 4 Introduces Multi Frame Generation & Enhancements For All DLSS Technologies". Nvidia. Retrieved January 7, 2025.
- ^ Mujtaba, Hassan (January 6, 2025). "Nvidia DLSS 4 Delivers An Insane 8x Performance Boost Versus DLSS 3 With Multi Frame Generation Technology, Enhanced Upscaling For RTX 20 & Above". Wccftech. Retrieved January 7, 2025.
- ^ "VESA to Update DisplayPort 2.1 with New Active Cable Specification for Up to 3x Longer DP80 Cables". VESA (Press release). January 6, 2025. Retrieved January 6, 2025.