Nvidia RTX 5050 puts Blackwell within reach of more gamers at $249 — entry-level 50-series launches in late July
How did your country report this? Share your view in the comments.
Diverging Reports Breakdown
Nvidia launches $249 RTX 5050, putting Blackwell within reach of more gamers — entry-level 50-series arrives in late July
N Nvidia has unveiled the desktop and notebook versions of the RTX 5050. The entry-level Blackwell card launches in the second half of July, starting at $249. It features 2,560 Blackwell CUDA cores, 8GB of GDDR6 memory running on a 128-bit wide interface, and a 130W TGP. Boost clocks are rated at up to 2,570MHz, but we can expect some AIB partner models to sport factory overclocks with even higher clock speeds. For 20% more money, or $50 more, buyers can get the RTX5060, which has 50% more CUDACore and significantly more memory bandwidth.
The RTX 5050 features 2,560 Blackwell CUDA cores, 8GB of GDDR6 memory running on a 128-bit wide interface, and a 130W TGP (which will necessitate a supplementary 6-pin or 8-pin connector). Boost clocks are rated at up to 2,570MHz, but we can expect some AIB partner models to sport factory overclocks with even higher clock speeds. The GPU’s 5th-gen tensor cores are rated for up to 421 TOPS of AI performance, and 40 TFLOPs from the GPU’s ray tracing cores.
Nvidia’s RTX 5050 is the first xx50-class desktop card since the RTX 3050. Nvidia never made an RTX 4050 for desktops. Despite being two generations apart, the RTX 5050 shares the same core specifications as the RTX 3050 8GB, including CUDA core count, 8GB memory capacity, and TGP.
(Image credit: Nvidia)
As part of its claimed performance boost over the RTX 3050, the RTX 5050 leans heavily on DLSS Multi-Frame Generation in titles that support it—it’s the first xx50-series desktop card that supports this technology. For games that don’t, the improvements from the two-generation newer Blackwell architecture and substantially higher clock speeds will have to compensate for the lack of improvement elsewhere.
The RTX 5050’s specs represent a significant reduction in compute power compared to its bigger brother the RTX 5060. The RTX 5060 boasts 50% more CUDA cores and around 40% more memory bandwidth thanks to its GDDR7 memory modules. For what we presume are cost-saving purposes, Nvidia has opted to give the RTX 5050 GDDR6 memory, making it the only Blackwell GPU so far to feature GDDR6. Even the mobile RTX 5050 comes with GDDR7 memory.
At its $249 MSRP, the RTX 5050 lands close to some stiff competition. RTX 5060 cards are available for $299 at retail, and AMD’s RX 9060 8GB can also be found for $299. Intel’s Arc B570 is selling for $279.99 right now, and features 2GB more memory. Just like Nvidia, AMD hasn’t introduced an entry-level GPU since the RX 6500 XT, but we might now expect a Radeon competitor to the RTX 5050 eventually. (AMD prototyped an RX 7500 6GB but never released it.)
You’ll have to wait for our full review of the RTX 5050 to see our full thoughts, but at first glance, the RTX 5050’s $249 MSRP may not be a great value, just like so many xx50-class GPUs of the past. For 20% more money, or $50 more, buyers can get the RTX 5060, which has 50% more CUDA cores and significantly more memory bandwidth. We’ll see how those differences play out when we put the RTX 5050 to the test.