NVIDIA Corp. unveiled its latest lineup of PC graphics cards as well as a new host of gaming-centric offerings at a GeForce special event on Sept. 1.

Appearing on the live-streamed event from his kitchen, the company's CEO Jensen Huang introduced the new lineup of RTX 30-series graphics processing units, or GPUs. Powered by NVIDIA's new "Ampere" architecture, the new consumer-oriented GPUs succeed the company's RTX 20-series "Turing" architecture hardware, which debuted in fall 2018.

"The generational leap is ultimately the most important factor of new GPUs," Huang said. "Ampere is the biggest generational leap we've had. [It] is an incredible two times the performance and energy efficiency of Turing."

The RTX 3080, which Huang called the flagship GPU of the new lineup, will be released Sept. 17 at a starting price of $699. With 10 GB of the new high-speed GDDR6X discrete graphics memory solution from Micron Technology Inc., the GPU will deliver "twice the performance of the [previous generation's] RTX 2080 at the same price," the CEO added.

The lower-end model of the lineup, the RTX 3070, will launch in October at a starting price of $499. It has 8 GB of GDDR6 memory and is faster than the previous generation's RTX 2080 Ti, which still retails for around $1,200.

Rounding out the lineup is the RTX 3090, which NVIDIA has dubbed the "BFGPU," or big ferocious GPU. Set to launch Sept. 24 at a starting price of $1,499, it will have 24 GB of GDDR6X memory and will include a silencer to make it up to 50% faster and 10x quieter than the previous generation's top-tier Titan RTX.

The 3090 model will allow users to play games at 60 frames per second in 8K resolution, Huang said.

All three GPUs have a dual-axial, flow-through thermal solution that provides up to 2x more cooling performance than the previous-generation hardware. The GPUs also support HDMI 2.1 functionality, allowing a single cable connection to 8K HDR TVs for ultra-high-resolution gaming.

Huang also introduced a range of new features to go along with the new lineup of GPUs, including NVIDIA Reflex, a suite of technologies that measure and optimize system latency to improve the performance of competitive video games. Set to release in September, these include NVIDIA Reflex Low-Latency Mode, which will be integrated into esports games such as Electronic Arts Inc.'s "Apex Legends," Activision Blizzard Inc.'s "Call of Duty: Warzone" and Epic Games Inc.'s "Fortnite" to reduce latency by up to 50%. Meanwhile, the NVIDIA Reflex Latency Analyzer detects input coming from the mouse and measures the time it takes for the resulting pixels to change on-screen.

Also launching in September is NVIDIA Broadcast, an app that will allow video game live streamers to upgrade any room into a home broadcast studio by transforming standard webcams and microphones into smart devices through AI-powered features, such as audio noise removal, virtual background effects and webcam auto frame.

Huang also previewed NVIDIA Omniverse Machinima, an app built onto the company's Omniverse 3D workflow collaboration platform that will allow users to take assets from supported games and use their webcams and AI to create characters, add high-fidelity physics and face and voice animation, to publish film-quality cinematics. The beta testing for the app will start in October.