NVIDIA GeForce RTX 4070 Ti – Enthusiast desktop GPU

Enthusiast-class desktop graphics card based on the Ada Lovelace AD104-400 GPU with 7,680 CUDA cores, 240 4th-gen Tensor Cores and 60 3rd-gen RT Cores for high-refresh 1440p and 4K gaming. Base clock 2310 MHz with boost up to 2610 MHz, 12 GB GDDR6X memory on a 192-bit bus delivering 504.2 GB/s bandwidth, 285 W TDP, PCIe 4.0 x16 interface and dual-slot design with 1x HDMI 2.1 and 3x DisplayPort 1.4a outputs.

Compare

Description

The NVIDIA GeForce RTX 4070 Ti is an enthusiast-class desktop graphics card based on the Ada Lovelace AD104-400-A1 GPU, launched in January 2023 as part of the GeForce RTX 40 series. Built on TSMC’s 5 nm process with 35.8 billion transistors and a 294 mm² die, it targets gamers who want high-refresh 1440p and capable 4K performance with full DirectX 12 Ultimate support. At launch, it carried a price of 799 USD and replaced high-end Ampere cards such as the GeForce RTX 3080 Ti and RTX 3090 in many gaming desktops [page:1].

The AD104 GPU powering the RTX 4070 Ti integrates 7,680 shading units (CUDA cores), 240 texture mapping units and 80 render output units, along with 240 fourth-generation Tensor Cores and 60 third-generation RT Cores for AI workloads and hardware-accelerated ray tracing. The GPU operates at a base frequency of 2310 MHz and can boost up to 2610 MHz under sufficient thermal and power headroom. This configuration delivers up to 40.09 TFLOPS of FP32 compute performance, 40.09 TFLOPS of FP16 at a 1:1 ratio, and high pixel and texture fill rates of 208.8 GPixel/s and 626.4 GTexel/s, respectively, making it ideal for demanding modern titles at high resolutions [page:1].

NVIDIA pairs the GeForce RTX 4070 Ti with 12 GB of GDDR6X memory connected via a 192-bit interface. The memory runs at an effective 21 Gbps data rate (1313 MHz actual), providing 504.2 GB/s of memory bandwidth to feed the GPU in high-resolution gaming and content creation workloads. The card also features 48 MB of L2 cache and 128 KB of L1 cache per SM, significantly improving effective bandwidth and latency for complex scenes and compute shaders compared to previous generations [page:1].

Board design specifications list the RTX 4070 Ti as a dual-slot card with dimensions of 285 mm in length, 112 mm in height and 42 mm in width, using a cooling solution optimized for high-end gaming systems. The card has a rated power draw of 285 W TDP and draws power from a single 16-pin PCIe connector, with NVIDIA recommending a 600 W power supply for stable operation. It connects to the rest of the system using a PCI Express 4.0 x16 interface, offering ample bandwidth for current and upcoming games and GPU-accelerated applications [page:1].

Display outputs on the GeForce RTX 4070 Ti include one HDMI 2.1 port and three DisplayPort 1.4a connectors, supporting multi-monitor setups and high-refresh 4K displays with variable refresh rate technologies. As an Ada Lovelace GPU, it supports DirectX 12 Ultimate, OpenGL 4.6, Vulkan 1.4 and CUDA 8.9, providing full access to hardware-accelerated ray tracing, DLSS, advanced shading features and a mature CUDA compute ecosystem. The GPU also integrates 8th-generation NVENC and 5th-generation NVDEC engines for efficient video encoding and decoding [page:1].

Overall, the NVIDIA GeForce RTX 4070 Ti delivers strong enthusiast-level performance for gamers and creators who want fast 1440p and 4K gaming, advanced ray tracing and DLSS support, and a modern feature set in a relatively efficient 285 W dual-slot design, making it a compelling choice for high-end gaming PC builds and upgrades [page:1].

General

GPU Name AD104
GPU Variant AD104-400-A1
Foundry TSMC
Process Size 5 nm
Die Size 294 mm²
Transistors 35.800 million
Density 121.8M / mm²

Architecture

GPU Architecture Ada Lovelace
Generation GeForce 40
Shading Units 7680
TMUs 240
ROPs 80
SM Count 60
CUDA 8.9
Tensor Cores 240
RT Cores 60

Frequency

GPU Base Clock 2310 MHz
GPU Boost Clock 2610 MHz
Memory Clock 1313 MHz 21 Gbps effective

Memory

Memory Size 12 GB
Memory Type GDDR6X
Memory Bus 192 bit
Bandwidth 504.2 GB/s

Cache

L1 Cache GPU 128 KB (per SM)
L2 Cache GPU 48 MB

Performance

Pixel Rate 208.8 GPixel/s
Texture Rate 626.4 GTexel/s

APIs & Compatibility

DirectX 12 Ultimate(12_2)
OpenGL 4.6
OpenCL 3.0
Vulkan 1.4
Shader Model 6.8
NVENC 8th Gen
NVDEC 5th Gen

Physical & Power

PCIe Interface PCIe 4.0 x16
Slot Width Dual-slot
Length 285 mm 11.2 inches
Height 112 mm 4.4 inches
TDP GPU 285 W
Suggested PSU 600 W
Power Connectors 1x 16-pin
Outputs 1x HDMI 2.13x DisplayPort 1.4a