NVIDIA GeForce RTX 4090 Founders Edition (24GB)

The NVIDIA GeForce RTX 4090 Founders Edition remains one of the most powerful graphics cards available. Built on NVIDIA's groundbreaking Ada Lovelace architecture, this flagship GPU redefined the possibilities of single-card gaming, professional content creation, and accelerated computing. Even with the arrival of NVIDIA's newer Blackwell generation, the RTX 4090 continues to deliver exceptional performance that rivals the best GPUs on the market.

Designed for enthusiasts seeking uncompromising 4K gaming performance, creators working with demanding 8K video and complex 3D environments, and developers exploring local AI workloads, the RTX 4090 combines extraordinary processing power, generous video memory, and advanced AI acceleration. With pricing that has become considerably more accessible since launch, it remains a compelling choice for users who demand flagship-level capabilities.

Key Features

• 16,384 CUDA Cores – Exceptional Parallel Processing Power: Powered by NVIDIA's Ada Lovelace architecture and manufactured using a custom TSMC 4N process, the RTX 4090 delivers immense computational throughput, remaining a leader in raw shader performance among pre-Blackwell GPUs.

• 24GB GDDR6X Memory with 384-bit Interface: Operating at 21Gbps and delivering over 1TB/s of memory bandwidth, the substantial 24GB framebuffer handles demanding 4K and 8K textures, expansive open-world environments, complex creative projects, and local AI models with billions of parameters.

• 72MB L2 Cache – Optimized Memory Efficiency: A large on-chip cache minimizes memory-access bottlenecks, improving data availability and processing efficiency across gaming, rendering, and compute-intensive applications.

• 512 Fourth-Generation Tensor Cores – Advanced AI Acceleration: Delivering up to 1,321 AI TOPS, these dedicated AI processors power NVIDIA DLSS 3 Frame Generation while accelerating generative AI workloads, including Stable Diffusion, large language models (LLMs), and other machine-learning applications.

• 128 Third-Generation RT Cores – Realistic Ray Tracing: Offering up to 191 TFLOPS of ray-tracing performance, the RTX 4090 enables sophisticated lighting, reflections, and shadows in graphically demanding titles, including fully path-traced experiences such as Cyberpunk 2077 and Alan Wake 2.

• 82.6 TFLOPS FP32 Compute Performance: Exceptional single-precision processing capability represents a major generational advancement over NVIDIA's previous RTX 30-series flagship, delivering outstanding performance in graphics rendering and computational workloads.

• 2.52GHz Boost Clock – High-Speed GPU Performance: Impressive factory boost frequencies ensure exceptional out-of-the-box responsiveness, with additional overclocking potential available on both Founders Edition and compatible partner-designed graphics cards.

• AV1 Hardware Encoding and Decoding – Professional Video Processing: Dedicated hardware acceleration enables efficient, high-quality video compression, faster encoding workflows, and improved streaming quality compared with older codecs, making the RTX 4090 an excellent choice for video creators and broadcasters.
  

Specifications

• Architecture: NVIDIA Ada Lovelace, TSMC 4N
• CUDA Cores: 16,384
• Tensor Cores: 512 (4th-generation)
• RT Cores: 128 (3rd-generation)
• Memory: 24GB GDDR6X, 384-bit bus, 21Gbps, 72MB L2 cache
• Memory Bandwidth: ~1,008GB/s
• FP32 Compute: 82.6 TFLOPS
• AI Performance: Up to 1,321 AI TOPS
• Boost Clock: 2.52GHz
• Total Graphics Power: 450W (transient spikes up to 600W)
• Interface: PCIe 4.0 x16
• Display Outputs: DisplayPort 1.4a, HDMI 2.1a
• Recommended PSU: 850W
• Launch: October 2022, original MSRP $1,599.00