Nvidia Corp. today debuted a new flagship line of consumer graphics cards, the RTX 30 series, headlined by a 28-billion transistor chip touted as the “fastest GPU ever built for gaming and creative types.”

Besides being a major chip supplier to data centers, Nvidia is also the top maker of graphics processing units for playing video games. The company’s gaming GPU business generated revenues of $1.65 billion in Nvidia’s most recent quarter after growing 26% year-over-year. 

Nvidia’s new RTX 30 chips are based on its latest Ampere architecture, the same architecture that powers its top-of-the-line A100 artificial intelligence GPU for data centers. The chips are also made using a newer fabrication technology than the company’s earlier flagship consumer silicon. Nvidia has switched from a 16-nanometer manufacturing process to Samsung Electronics Co. Ltd.’s eight-nanometer process, leading to smaller transistors that can fit in larger numbers on each chip.

The company promises these improvements will produce a big jump in performance. According to Nvidia, the RTX 30 series offers up to twice the performance and 1.9 times the power efficiency of its previous-generation flagship consumer GPUs.

Headlining the RTX 30 line is the RTX 3090. The chip packs 28 billion transistors that can run video games at 60 frames per second with 8K resolution, which Nvidia says is a world first. Displays with 8K resolution are only starting to hit the market and offer 16 and four times the resolution of mainstream HD and 4K screens, respectively.

“There is clearly a need for a giant GPU that is available all over the world,” Nvidia Chief Executive Officer Jensen Huang (pictured) said during the virtual event announcing the chips today. “So, we made a giant Ampere.” Speaking about the RTX 30 line more generally, Huang described the series as the “greatest generational leap in company history.” 

The line includes two other chips on launch, the 3080 and 3070 GPUs, that will become available at more accessible price points than the RTX 3090. While the latter GPU is set to hit the market for $1,499, the 3080 will be available from $699, while the 3070 is set to carry a $499 price tag.

All three chips benefit from the enhancements in Nvidia’s Ampere architecture. Among them is the company’s third-generation Tensor Cores, specialized circuits optimized for AI, that have twice the throughput of the second generation. These AI circuits will enable the chips to more efficiently run Nvidia’s deep learning super sampling technology, which uses neural networks to increase the quality of video game visuals.

The chips also feature a new variety of memory supplied by Micron Technology Inc. that Nvidia touts as the “fastest discrete graphics memory” in the industry.

Photo: Nvidia