In what can only be described as a landmark moment for gaming and graphics technology, NVIDIA unveiled its latest gem at CES 2025: the RTX 50-series GPUs, with the spotlight shining brightly on the RTX 5090 and RTX 5080. The announcement was made by NVIDIA’s own Jensen Huang during the keynote.

The high-end marvels

The RTX 5090, the crown jewel of this series, comes with a hefty price tag of $1,999 but promises unprecedented performance. With 32GB of GDDR7 memory, a memory bandwidth of 1,792GB/sec, and a staggering 21,760 CUDA cores, NVIDIA claims this beast is not just an evolution but a revolution, boasting performance that’s twice as fast as its predecessor, the RTX 4090. The RTX 5090 leverages NVIDIA’s new Blackwell architecture and DLSS 4 technology for these gains.

However, this power does come with a caveat: a total graphics power (TGP) of 575 watts, necessitating a 1000-watt power supply. This is a significant jump from the RTX 4090’s power consumption, yet NVIDIA assures that efficiency improvements mean you won’t often max out this power envelope.

Live demonstrations at CES included running “Cyberpunk 2077” where the RTX 5090 with DLSS 4 achieved a frame rate of 244 fps, compared to 112 fps on the RTX 4090 using DLSS 3.5, both with full ray tracing enabled, showcasing the raw power of the new card.

The RTX 5080, priced at $999, isn’t far behind, promising to double the performance of the RTX 4080. It sports 16GB of GDDR7 memory, 960GB/sec bandwidth, and 10,752 CUDA cores, with a TGP of 360 watts, suggesting a recommended PSU of 850 watts.

The mid-range contenders

Not to be outdone, NVIDIA also introduced the RTX 5070 Ti and RTX 5070. The RTX 5070 Ti, at $749, features 16GB of GDDR7 memory, 896GB/s bandwidth, and 8,960 CUDA cores, with a TGP of 300 watts. Meanwhile, the RTX 5070, priced at $549, comes with 12GB of GDDR7 memory, 672 GB/sec bandwidth, and 6,144 CUDA cores, with a TGP of 250 watts. These cards aim to democratize high-end gaming performance, with Huang claiming the RTX 5070 offers “RTX 4090 performance at $549,” largely thanks to advancements in DLSS.

Laptop lineup

NVIDIA isn’t stopping at desktops; the RTX 50-series will also grace laptops. The RTX 5090 for laptops features 24GB of GDDR7, while the RTX 5080, RTX 5070 Ti, and RTX 5070 come with 16GB, 12GB, and 8GB respectively. These laptop GPUs are set to launch in March, providing mobile gamers with a taste of this high-performance architecture.

AI and graphics innovations

The keynote also delved into NVIDIA’s AI advancements. The RTX Neural Shaders and RTX Neural Faces, leveraging generative AI, aim to enhance textures and facial details in games. DLSS 4’s introduction was perhaps the most anticipated, with its Multi Frame Generation technology promising to boost frame rates by up to 8x through advanced AI prediction and rendering techniques. This upgrade isn’t exclusive to the new GPUs; existing RTX cards will also benefit from DLSS 4’s transformer-based models for improved image quality and reduced latency.

Looking forward

This announcement marks over two years since NVIDIA’s last major GPU launch, the RTX 40-series, which was centered around enhancing ray tracing with DLSS 3. The RTX 50-series not only continues this legacy but pushes the boundaries further with AI integration into graphics processing, setting a new standard for what gamers and professionals can expect from their graphics hardware.

As we gear up for the availability of these GPUs starting January 30th for the RTX 5090 and RTX 5080, with the others following in February, the gaming and content creation community are on the edge of their seats, eager to experience what NVIDIA describes as the future of graphics technology.