Explore canonical things, gear specs, and community sparks featuring NVIDIA.
Revered as one of the greatest value-for-performance graphics cards ever released, delivering near 8800 GTX performance at a mainstream price point on a 65nm G92 die.
Historic G80 unified shader architecture milestone with 768MB GDDR3, 384-bit bus, DirectX 10 support, and unmatched longevity.
The most popular GPU in Steam Hardware Survey history for over five consecutive years: 6GB 192-bit GDDR5, 1,280 CUDA cores, and 120W TDP.
Bestselling 8GB GDDR5 Pascal card with 1,920 CUDA cores, 150W TDP, and Titan X-beating performance in 2016.
High-performance Pascal card with 8GB GDDR5, 2,432 CUDA cores, and 180W TDP nearly matching the GTX 1080.
The historic 16nm FinFET breakthrough: 8GB 10 Gbps GDDR5X, 2,560 CUDA cores, faceted geometric magnesium casing, and radial blower fan.
The greatest generational leap in consumer GPU history: 11GB GDDR5X, 3,584 CUDA cores, 352-bit bus, and faceted magnesium blower shroud.
Massive budget 1080p champion powered by Turing architecture without RT/Tensor cores, supercharged with 6GB of fast 14Gbps GDDR6 VRAM.
Original GF100 Fermi architecture GPU featuring direct contact copper heatpipes protruding through the heat sink, 480 CUDA cores, and legendary heat output.
Refined GF110 Fermi flagship with fully unlocked 512 CUDA cores, vapor chamber cooling, and significant thermal improvements over the GTX 480.
First 28nm Kepler GPU introducing GPU Boost automatic overclocking, SMX streaming multiprocessors, and unprecedented power efficiency.
The fully unlocked Kepler GK110 monster featuring 2880 CUDA cores, 3GB of 7GHz GDDR5 memory, and iconic magnesium and aluminum reference blower cooler.
The definitive gaming card of the mid-2010s: 4GB (3.5GB + 0.5GB segmented) GDDR5, 1,664 CUDA cores, and incredible overclocking headroom at 145W TDP.
Maxwell flagship with full GM204 die, 4GB 7 Gbps GDDR5, 2,048 CUDA cores, and cast aluminum blower shroud.
Maxwell architecture icon with 6GB GDDR5, 2,816 CUDA cores, and vapor chamber cooling.
Original 2013 GK110 Kepler supercomputer GPU introducing the iconic cast-aluminum industrial design with 6GB GDDR5 and double-precision compute.
Entry point into real-time ray tracing and Tensor core DLSS AI upscaling, featuring 1920 CUDA cores and 6GB of GDDR6 memory.
Upgraded 8GB 256-bit GDDR6 mainstream Turing card with 2,176 CUDA cores and dual-fan silver aluminum shroud.
High-end Turing graphics card featuring 2304 CUDA cores, 8GB GDDR6 VRAM, and factory overclocked vapor chamber dual-fan aluminum shroud.
The defining price-to-performance card of the Turing generation with 8GB GDDR6, 2,560 CUDA cores, and RTX 2080 class gaming performance.
The pioneering Turing-generation desktop GPU that introduced hardware real-time ray tracing and DLSS to PC gaming, featuring 8GB GDDR6 and dual axial fans.
Turing graphics card with mirrored center badge, 8GB 15.5 Gbps GDDR6, 3,072 CUDA cores, and 250W TGP.
Landmark Turing flagship introducing real-time hardware ray tracing and DLSS with 11GB GDDR6, 4,352 CUDA cores, dual-axial fans, and full-length vapor chamber.
The dominant long-term leader on the Steam Hardware Survey, featuring 12GB of GDDR6 VRAM, 3584 CUDA cores, and 2nd generation RT cores for mainstream 1080p/1440p gaming.
8GB GDDR6 mid-range champion with 4,864 CUDA cores outperforming the RTX 2080 Super at 200W TGP.
8GB GDDR6 2-slot graphics card delivering RTX 2080 Ti performance at 220W TGP.
Upgraded GA104 Ampere GPU featuring 6144 CUDA cores and ultra-fast 8GB GDDR6X memory on a 256-bit bus with distinctive Flow-Through Founders Edition cooler.
Groundbreaking Ampere generation flagship with 8,704 CUDA cores, 10GB 19 Gbps GDDR6X, dual-axial flow-through cooling, and 320W TGP.
12GB GDDR6X graphics card with 10,240 CUDA cores, 384-bit memory bus, and 350W TGP in dual-slot flow-through shroud.
Massive 3-slot "BFGPU" with 24GB GDDR6X, 10,496 CUDA cores, NVLink SLI connector, and 350W TGP.
Uncapped Ampere flagship with full GA102 die (10,752 CUDA cores), 24GB 21 Gbps GDDR6X, PCIe 5.0 16-pin 12VHPWR connector, and 450W TGP.
Entry Ada Lovelace card with 8GB GDDR6, 3,072 CUDA cores, 8-lane PCIe 4.0 interface, and 115W TGP.
Mainstream 1080p and 1440p gaming and AI card with expanded 16GB GDDR6 VRAM, 4,352 CUDA cores, and ultra-efficient 165W TGP.
Mainstream 8GB GDDR6 graphics card with 4,352 CUDA cores, DLSS 3 Frame Generation, and 160W TGP.
12GB GDDR6X compact 2-slot Founders Edition card delivering RTX 3080 performance at an efficient 200W TGP.
12GB GDDR6X graphics card with 7,168 CUDA cores (20% more than RTX 4070), blackout shroud, and 220W TGP.
16GB GDDR6X Ada Lovelace graphics card with 9,728 CUDA cores, dual axial flow-through cooling, and DLSS 3 support.
16GB GDDR6X graphics card in full matte-black Founders Edition shroud with 10,240 CUDA cores, 23 Gbps memory speed, and 320W TGP.
Flagship 24GB GDDR6X Ada Lovelace graphics card with 16,384 CUDA cores, 3rd Gen RT cores, 4th Gen Tensor cores, DLSS 3 Frame Generation, and 450W TGP.
NVIDIA SHIELD TV Pro premium streaming media player with Tegra X1+ processor, AI-enhanced 4K upscaling, Dolby Vision and Dolby Atmos, 3GB RAM, and 16GB storage.
Explore canonical things, gear specs, and community sparks featuring NVIDIA.
Revered as one of the greatest value-for-performance graphics cards ever released, delivering near 8800 GTX performance at a mainstream price point on a 65nm G92 die.
Historic G80 unified shader architecture milestone with 768MB GDDR3, 384-bit bus, DirectX 10 support, and unmatched longevity.
The most popular GPU in Steam Hardware Survey history for over five consecutive years: 6GB 192-bit GDDR5, 1,280 CUDA cores, and 120W TDP.
Bestselling 8GB GDDR5 Pascal card with 1,920 CUDA cores, 150W TDP, and Titan X-beating performance in 2016.
High-performance Pascal card with 8GB GDDR5, 2,432 CUDA cores, and 180W TDP nearly matching the GTX 1080.
The historic 16nm FinFET breakthrough: 8GB 10 Gbps GDDR5X, 2,560 CUDA cores, faceted geometric magnesium casing, and radial blower fan.
The greatest generational leap in consumer GPU history: 11GB GDDR5X, 3,584 CUDA cores, 352-bit bus, and faceted magnesium blower shroud.
Massive budget 1080p champion powered by Turing architecture without RT/Tensor cores, supercharged with 6GB of fast 14Gbps GDDR6 VRAM.
Original GF100 Fermi architecture GPU featuring direct contact copper heatpipes protruding through the heat sink, 480 CUDA cores, and legendary heat output.
Refined GF110 Fermi flagship with fully unlocked 512 CUDA cores, vapor chamber cooling, and significant thermal improvements over the GTX 480.
First 28nm Kepler GPU introducing GPU Boost automatic overclocking, SMX streaming multiprocessors, and unprecedented power efficiency.
The fully unlocked Kepler GK110 monster featuring 2880 CUDA cores, 3GB of 7GHz GDDR5 memory, and iconic magnesium and aluminum reference blower cooler.
The definitive gaming card of the mid-2010s: 4GB (3.5GB + 0.5GB segmented) GDDR5, 1,664 CUDA cores, and incredible overclocking headroom at 145W TDP.
Maxwell flagship with full GM204 die, 4GB 7 Gbps GDDR5, 2,048 CUDA cores, and cast aluminum blower shroud.
Maxwell architecture icon with 6GB GDDR5, 2,816 CUDA cores, and vapor chamber cooling.
Original 2013 GK110 Kepler supercomputer GPU introducing the iconic cast-aluminum industrial design with 6GB GDDR5 and double-precision compute.
Entry point into real-time ray tracing and Tensor core DLSS AI upscaling, featuring 1920 CUDA cores and 6GB of GDDR6 memory.
Upgraded 8GB 256-bit GDDR6 mainstream Turing card with 2,176 CUDA cores and dual-fan silver aluminum shroud.
High-end Turing graphics card featuring 2304 CUDA cores, 8GB GDDR6 VRAM, and factory overclocked vapor chamber dual-fan aluminum shroud.
The defining price-to-performance card of the Turing generation with 8GB GDDR6, 2,560 CUDA cores, and RTX 2080 class gaming performance.
The pioneering Turing-generation desktop GPU that introduced hardware real-time ray tracing and DLSS to PC gaming, featuring 8GB GDDR6 and dual axial fans.
Turing graphics card with mirrored center badge, 8GB 15.5 Gbps GDDR6, 3,072 CUDA cores, and 250W TGP.
Landmark Turing flagship introducing real-time hardware ray tracing and DLSS with 11GB GDDR6, 4,352 CUDA cores, dual-axial fans, and full-length vapor chamber.
The dominant long-term leader on the Steam Hardware Survey, featuring 12GB of GDDR6 VRAM, 3584 CUDA cores, and 2nd generation RT cores for mainstream 1080p/1440p gaming.
8GB GDDR6 mid-range champion with 4,864 CUDA cores outperforming the RTX 2080 Super at 200W TGP.
8GB GDDR6 2-slot graphics card delivering RTX 2080 Ti performance at 220W TGP.
Upgraded GA104 Ampere GPU featuring 6144 CUDA cores and ultra-fast 8GB GDDR6X memory on a 256-bit bus with distinctive Flow-Through Founders Edition cooler.
Groundbreaking Ampere generation flagship with 8,704 CUDA cores, 10GB 19 Gbps GDDR6X, dual-axial flow-through cooling, and 320W TGP.
12GB GDDR6X graphics card with 10,240 CUDA cores, 384-bit memory bus, and 350W TGP in dual-slot flow-through shroud.
Massive 3-slot "BFGPU" with 24GB GDDR6X, 10,496 CUDA cores, NVLink SLI connector, and 350W TGP.
Uncapped Ampere flagship with full GA102 die (10,752 CUDA cores), 24GB 21 Gbps GDDR6X, PCIe 5.0 16-pin 12VHPWR connector, and 450W TGP.
Entry Ada Lovelace card with 8GB GDDR6, 3,072 CUDA cores, 8-lane PCIe 4.0 interface, and 115W TGP.
Mainstream 1080p and 1440p gaming and AI card with expanded 16GB GDDR6 VRAM, 4,352 CUDA cores, and ultra-efficient 165W TGP.
Mainstream 8GB GDDR6 graphics card with 4,352 CUDA cores, DLSS 3 Frame Generation, and 160W TGP.
12GB GDDR6X compact 2-slot Founders Edition card delivering RTX 3080 performance at an efficient 200W TGP.
12GB GDDR6X graphics card with 7,168 CUDA cores (20% more than RTX 4070), blackout shroud, and 220W TGP.
16GB GDDR6X Ada Lovelace graphics card with 9,728 CUDA cores, dual axial flow-through cooling, and DLSS 3 support.
16GB GDDR6X graphics card in full matte-black Founders Edition shroud with 10,240 CUDA cores, 23 Gbps memory speed, and 320W TGP.
Flagship 24GB GDDR6X Ada Lovelace graphics card with 16,384 CUDA cores, 3rd Gen RT cores, 4th Gen Tensor cores, DLSS 3 Frame Generation, and 450W TGP.
NVIDIA SHIELD TV Pro premium streaming media player with Tegra X1+ processor, AI-enhanced 4K upscaling, Dolby Vision and Dolby Atmos, 3GB RAM, and 16GB storage.