These days a multitude of rumors of the RTX 4090 Ti or RTX Titan Ada, as the devil is going to call NVIDIA at the end. From assemblies of what the graphics card will be like, originating from sources with zero credibility, to technical specifications. However, we believe that it is too early to release this graphics card and that it does not make sense for Jen Hsen Huang’s people. At least for the moment.
When it comes to raw power, the RTX 4090 is pretty wild stuff, and we’re not just saying that, even the most powerful processors from Intel and AMD, the Intel Core i9-13900KS and Ryzen 9 7950X are a bottleneck. That is, they do not allow 100% of this powerful graphics card to be removed. This being one of the reasons why we believe that an RTX 4090 Ti should not be a priority for NVIDIA, especially if we take into account the current situation.
There will be RTX 4090 Ti, but not when or how you expect
It took NVIDIA no more and no less than 18 months to launch the RTX 3090 Ti compared to the launch of the original RTX 3090. Before that, they dedicated themselves to filling the market with different models, designed for different ranges of consumers and always with an orderly and staggered launch in terms of the different ranges. At the moment with the RTX 40 the same modus operandi is being repeated.
In any case, a detail that no one has noticed, completely obscured by gossip on the net, is that the reason why the industrial design of the RTX 4090 is higher than that of the RTX 3090 Ti, despite having chips of the same size and consuming the same, is to give room for the launch of the 600 W version. That is, the RTX 4090 Ti and cut costs. Thus, the standard and Ti versions of each of the RTX 40 share common elements in order to cut production costs and speed time to market.
So, logically, we should not expect a different industrial design for the RTX 4090 Ti compared to the current model, nor a change in the PCB of the card. Rather the use of a version of the AD102 chip with better specs and faster GDDR6X. Specifications that have recently appeared.
What do we know, for the moment, of its technical specifications?
Officially and for the moment nothing, but the latest rumors point to the following specifications:
- The power required by the graphics card will be 600W, the maximum that the 12+4 pin 12VHPWR cable can give in its official specification. Hence forget about 800 W configurationswhich was already nonsense from the first moment.
- Video memory will continue to be 24GB GDDR6Xbut with a speed of 24 Gbps instead of the 21 Gbps of the RTX 4090 current. This would increase the bandwidth at 1152 GB/s.
- The number of nuclei or Active SM will be 142, with 18,766 active 32-bit floating point units. The final clock speed is unknown.
In any case, there may still be last-minute changes and we are not expecting them on the market yet, due to what we have told you a few lines ago. There is still no desktop CPU that does not limit the power of the RTX 4090, so it is pointless to launch a much more powerful card. There is also no incentive to launch an even more expensive one. It is for all this that we think that its launch will be similar to that of the RTX 3090 Ti and that we should expect it rather by 2024, coinciding with the next generation of Intel and AMD processors.