NVIDIA is not going to leave anything to chance with the new RTX 40 and according to what has been leaked, Huang’s bet will be to start with the best they have to unseat AMD and incidentally allow a second strategy that seen what has been seen will be necessary given the current context in which we move. Therefore, the first graphics card to reach the market within the new RTX 40 it will be the RTX 4090, why does NVIDIA change its movement?
Those in green are going to change the formula for success that had worked for them until now to do something totally different from the previous one, surely forced by something we talked about this morning and that we will link to below, since according to the schedule of launch of the RTX 40 is already set in advance.
The first RTX 40 will be the most expensive… Out of obligation
There will be no high or high mid-range at launch, although there will be in the presentation as such and this should be distinguished because it is important for what we are going to deal with. NVIDIA will present three graphics cards and the new Ada Lovelace architecture at some point in July. Specifically, the three models will be the RTX 4090, RTX 4080 and RTX 4070 as expected, but unlike other generations, only the first one will be launched, this being the top of the range for at least a year and a half.
nVidia’s launch strategy for “Ada Lovelace”: 2 paths seem possible.
A) 4080 → 4090 → 4070
B) 4090 → 4080 → 4070
— 3DCenter.org (@3DCenter_org) May 29, 2022
The release order is also important here because it draws the line that NVIDIA wants to follow almost out of obligation, and we explain ourselves. According to a famous leaker as it is Kopite7kimi will come first RTX 4090after the RTX4080 and finally the RTX4070followed later and without specific dates or models of the mid and low range of RTX 40 graphics cards.
The questions then crowd us, why does NVIDIA change the launch method? The answer is complex, but interesting.
Three chips for three different GPUs and the RTX 30 in stores
First of all, NVIDIA knows that manufacturers and retailers have a huge stock of graphics cards that people aren’t buying because they’re waiting for the new RTX 40s. RTX 4070 you leave months in between where stores can offer discounts while the market stabilizes.
Risky bet, but certainly not silly. In an inflationary market like this where the user has been mistreated and where, as we have seen this morning, retailers continue to play at inflating prices to try to lose as little money as possible after purchasing from manufacturers, perhaps NVIDIA is playing a game that It can be expensive if AMD steps on the accelerator.
On the other hand, we have to talk about the volume of the TSMC 4N node and the chips. This node is exclusive to NVIDIA and because of this certain concessions are made. Those of Huang have arranged three different chips for the three models of graphics cards named:
- RTX 4090 -> AD102-300
- RTX 4080 -> AD103S
- RTX 4070 -> AD104-400
Or at least that’s the theory according to the most recent leaks. The temporary segmentation gives TSMC a break and allows the defective AD102-300 chips to be reused and rebranded at least for the RTX 4080 as AD103S, or even launched as an alleged AD102-400 capped of performance. In any case and theories aside, the RTX 4090 should be presented before the end of August focused on playing freely in 4K and arrive at the latest in mid-September, while the RTX 4080 and RTX 4070 should do the same before the end of this 2022.