Nvidia RTX 50 Series revealed: Release date, prices, specifications, & more
NVIDIANvidia has finally lifted the veil on its highly-anticipated RTX 50 Series cards. From performance details to prices, here’s the rundown on all there is to know.
The RTX 50 Series of GPUs is finally upon us. After months of speculation, the new range of cards was unveiled at CES 2025 as Nvidia’s keynote turned heads.
The big driver? AI-powered rendering, as Nvidia pushed the power of artificial intelligence over all else in this wave of hardware. So if you’re wondering just what it all means and how to get your hands on one, here’s what we know.
RTX 50 series release dates
The RTX 50 Series will initially be available starting in January, 2025, with each card getting its own time in the sun. Below is a quick breakdown of key dates:
- GeForce RTX 5090 GPU: Available January 30, 2025
- GeForce RTX 5080 GPU: Available January 30, 2025
- GeForce RTX 5070 Ti GPU: Available February
- GeForce RTX 5070 GPU: Available February
Exactly how many units are being manufactured for this initial wave of orders is yet unclear, though as usual, it’d be wise to anticipate demand not being met. Much like the 50 Series, cards will certainly sell out quite quickly.
We’ll be sure to update you here with further pre-order details once they become available.
RTX 50 series prices
RTX 50 Series GPU prices were revealed at CES 2025. With four cards on offer, there are four different prices for the various tiers. Below is a quick look:
- GeForce RTX 5090: $1,999 USD
- GeForce RTX 5080: $999 USD
- GeForce RTX 5070 Ti: $749 USD
- GeForce RTX 5070: $549 USD
NvidiaRTX 50 series specifications
As you would expect, the 50 Series GPUs are a generational leap ahead of their predecessors, with the 5090 alone promising double the performance of the 4090.
| Card | RTX 5090 | RTX 5080 | RTX 5070 Ti | RTX 5070 |
| GPU Chip | GB202-300 | GB203-400 | GB203-300 | GB205-300 |
| Cuda Cores | 21760 | 10752 | 8960 | 6144 |
| Memory | 32 GB GDDR7 | 16GB GDDR7 | 16GB GDDR7 | 12 GB GDDR7 |
| Max memory bus | 512-bit | 256-bit | 256-bit | 192-bit |
| Memory speed | 28Gbps | 30 Gbps | 28 GBps | 28 GBps |
| Memory Bandwidth | 1792 GB/s | 960 GB/s | 896 GB/s | 672 GB/s |
Wattage
| GPU | TDP (W) |
| RTX 5090 | 575W |
| RTX 5080 | 360W |
| RTX 5070 Ti | 300W |
| RTX 5070 | 250W |
RTX 50 series features
As mentioned, the main force of the 50 Series cards is its utilization of AI, namely the Blackwell system boosting all other AI elements.
DLSS4, for instance, gets Multi Frame Generation, helping provide up to three distinct frames per rendered frame in other to smoothen out the image. Furthermore, DLSS 4 also brings significant enhancements to the likes of Ray Reconstruction and Super Resolution, as Nvidia outlined.
Meanwhile, Nvidia Reflex 2 helps keep games sharp despite the added frame generation, reducing latency by up to 75% for competitive games.