Because the RTX 2080 Ti wasn’t expensive enough or powerful enough in the first place, Nvidia has launched their latest Titan card the Titan RTX.

The Titan range is a weird consumer/prosumer crossover card that should be used by people that can’t afford or don’t need a Quadro/Tesla but ends up being used more frequently in excessive gamer builds by people with more money than sense. I won’t lie, I want one myself, and I would probably by one if I won the Euromillions. I could build my current PC two times over for the money of this card.

Based on the same TU102 Turing GPU as the GeForce RTX 2080 Ti and Quadro RTX 6000, the Titan RTX borrows most of its specification from the workstation-oriented Quadro RTX 6000: The card boasts six processing clusters with a total of 36 texture processing clusters for 72 streaming multiprocessors, 4,608 active CUDA cores, 758 Tensor cores, and 72 RT cores, all running at a base clock of 1,350MHz and boost clock of 1,770MHz. The card also includes 24GB of GDDR6 memory – non-ECC, naturally – running at an effective 7,000MHz on a 384-bit bus for 672GB/s of bandwidth.

The Titan RTX is expected to offer around 10% more performance than an RTX 2080 Ti even though it costs twice as much.

The card is not available to 3rd party companies so there will be no fancy cooling systems and RGB integrated into the Titan apart from aftermarket mods.

It includes three DisplayPort 1.4, one HDMI, and one USB Type C connector with VirtualLink compatibility, along with two eight-pin PCI Express connectors. Nvidia is claiming a 280W thermal design profile (TDP), though it recommends a 650W minimum power supply. The dual-slot design includes dual fans over a vapour chamber cooler covered in diecast aluminium, while an NVLink bridge connector allows for two cards to be paired up in SLI in a single system.

Nvidia has confirmed that the boards will be available later this month, priced at an eye-watering £2,399 (inc. VAT).

Nvidia Founders Edition Graphics  Titan RTX Geforce RTX 2080 Ti Geforce RTX 2080 Geforce RTX 2070
Architecture Turing Turing Turing Turing
CUDA Cores 4,608 4,352 2944  2304
Ray Tracing Performance ???? 10 Gigarays 8 Gigarays 6 Gigarays
Base Clock
1350MHz 1350MHz 1510Mhz 1410MHz
Boost Clock 1770MHz 1635MHz 1800MHz 1710MHz
Memory Type GDDR6 GDDR6 GDDR6 GDDR6
Memory Capacity 24GB 11GB 8GB 8GB
Memory Speed 14Gbps 14Gbps 14Gbps 14Gbps
Memory Bandwidth 672GB/s 616GB/s 448GB/s 448GB/s
Memory Bus Size 384-bit 352-bit 256-bit 256-bit
SLI Via NVLink Via NVLink Via NVLink N/A