It's been a long time coming, with months of speculation and a few leaks along the way, but Nvidia revealed its next generation RTX 4080 and RTX 4090 GPUs at GTC 2022.

Largely holding up to expectations though light on serious detail—as you might expect for a first reveal—Nvidia's 4000-Series GPUs will push graphics processing to the next level with its new Ada Lovelace architecture.

Nvidia Reveals the RTX 4090

At Nvidia's GTC 2022 "Special Broadcast," Nvidia CEO Jensen Huang revealed the all-new RTX 4090 and RTX 4080 GPUs, both of which are built using Nvidia's new Ada Lovelace architecture.

The RTX 4090 will become the company's flagship GPU. The RTX 4090 will launch on 12 October 2022 and will retail for $1,599. Furthermore, the RTX 4090 will launch with 24GB GDDR6X memory, with Nvidia claiming that its new flagship will perform around two to four times faster than the previous top-of-the-line GPU, the RTX 3090 Ti.

In addition, the RTX 4090 will feature a phenomenal 16,384 CUDA cores, which is off the scale, and a boost clock up to 2.52GHz. All this comes with a significant power requirement, with the new RTX 4090 launching with a 450W TDP—the same as the RTX 3090 Ti, but with almost twice the memory and processing power.

Finally, the 24GB RTX 4090 will feature Nvidia's massively upgraded DLSS 3.0, which should greatly improve processing and functionality across the board.

Nvidia Also Reveals the RTX 4080

During the same talk, Nvidia also revealed two RTX 4080 models, neither of which were accompanied by a specific release date, though November is the target month.

• The RTX 4080 16GB will launch with 9,728 CUDA cores (around 750 more than the RTX 3080), a TDP of 320W, and a suggested retail price of $1,199.
• The RTX 4080 12GB will launch with 7,680 CUDA cores, a TDP of 285W, and a suggested retail price of $899.

It is thought that the lower-specced 12GB version of the RTX 4080 was originally going to be launched as the RTX 4070, but there is no official word on why this was changed or indeed when the actual RTX 4070 will launch (or what its specs will be).

Nvidia's 4000-Series GPUs Use New Ada Lovelace Architecture

Like the RTX 4090, both versions of the RTX 4080 are built using the Ada Lovelace architecture, which means all three cards will use Nvidia's DLSS 3.0, the latest and most powerful version of its Deep Learning Super Sampling technology.

The third-generation Ray Tracing cores will see significant boosts across the board, with Nvidia boasting that its new DLSS 3.0 can generate additional frames during rendering to boost visuals, apparently boosting overall game performance up two times over its previous version found on RTX 3000-Series GPUs.

Nvidia's RTX 4000-Series Finally Launches

It's not long until you can finally get your hands on an Nvidia 4000-Series GPU. The RTX 4090 will officially launch on 12 October, and all being well, the RTX 4080 GPUs will launch around one month later, in November 2022.

The successful launch of new Nvidia GPUs is important for the company, especially as many criticized the previous generation for being underpowered—which it looks like Nvidia has addressed with the 12GB GDDR6X minimum for the RTX 4080 and the whopping 24GB GDDR6X for the RTX 4090.

What remains to be seen now is how Nvidia's RTX 4000 GPUs perform in real-world tests. Although the hype surrounding the reveal of the new GPUs is understandable, Nvidia's presentation was surprisingly sparse on benchmarking information. Now, first-party benchmarking is always taken with a grain of salt anyway, but some figures would have been nice.