A part of me thought the rumors about Nvidia’s RTX 50-series GPUs had been fallacious. However the RTX 5090 is simply as insane as everybody stated it could be. Nvidia CEO Jensen Huang took to the stage at CES 2025 to formally kick off the present, saying the RTX 5090, which clocks in at a staggering $2,000.
The monstrous graphics card has been the subject of rumors and hypothesis for nicely over a 12 months, as the complete vary of Blackwell GPUs appeared to slipped out of a launch late final 12 months, as rumors as much as that time had advised. The RTX 5090 marks a $400 improve over the RTX 4090 we noticed within the earlier era.
Nevertheless, Nvidia lowered costs down the stack. The RTX 5080 clocks in on the similar value because the RTX 4080 Tremendous, whereas each the RTX 5070 Ti and RTX 5070 are $50 lower than their last-gen counterparts. Nvidia says that the RTX 5070 supplies efficiency on the extent of an RTX 4090, which Huang attributed to the brand new GDDR7 reminiscence, which can be obtainable in all 4 GPUs.
For the flagship RTX 5090, Nvidia says it’s twice as quick because the RTX 4090, which you’ll be able to see within the video above. Throughout extra video games, Nvidia shared efficiency estimates, which you’ll be able to see beneath.
As rumors advised, the RTX 5090 comes with 32G of GDDR7 reminiscence, and it packs an influence draw of 575 watts. Nvidia continues to be in a position to ship that a lot wattage over a single 16-pin connector, which it has used within the earlier two generations of graphics playing cards.
Additional down the stack is the RTX 5080, which is rather more tame by comparability. It comes with upgraded GDDR7 reminiscence, as nicely, nevertheless it sticks with 16GB, matching what was obtainable on last-gen’s RTX 4080 Super.
Get your weekly teardown of the tech behind PC gaming
Though the playing cards aren’t anticipated to reach for no less than a couple of months, Nvidia additionally revealed the RTX 5070 Ti and RTX 5070. Each use GDDR7 reminiscence, with 16GB for the RTX 5070 Ti and 12GB for the RTX 5070. The reminiscence is quicker, however that’s the identical capability that was obtainable on the RTX 4070 Super. Fortunately, I haven’t encountered a sport the place 12GB of VRAM is a severe situation but.
As is the case with the flagship RTX 5090, Nvidia claims all of its new RTX 50-series GPUs are twice as quick as their last-gen counterparts. You’ll be able to see the efficiency estimates Nvidia shared for the three extra playing cards above. The entire efficiency estimates Nvidia shared used Nvidia’s new DLSS 4, which incorporates multi-frame era.
Huang confirmed off a brand new design for Nvidia’s Founder’s Version playing cards, as nicely. In contrast to the previous two generations, the RTX 50-series design comes with two followers on one facet of the graphics card, together with rounded corners and an angled 16-pin energy connector. As is normally the case, Nvidia is working with a number of board companions for inventory and overclocked fashions.
Nvidia says that the RTX 5090 and RTX 5080 can be obtainable on January 30, whereas the RTX 5070 Ti and RTX 5070 will begin rolling out in February. Along with desktop GPUs, Nvidia revealed RTX 50-series mobile graphics cards, that are anticipated to reach in March.
