NVIDIA GeForce RTX 4080 12GB should be the GeForce RTX 4070. Looks bad and bodes bad for future generations

During yesterday’s GeForce Beyond conference as part of GTC 2022, NVIDIA introduced a new GPU architecture – Ada Lovelace – and three graphics cards: GeForce RTX 4090, GeForce RTX 4080 16GB, and GeForce RTX 4080 12GB. While in the case of the top card we got what we expected for a long time (at least in terms of specs), in the case of the GeForce RTX 4080 the situation is at least strange and debatable. However, this isn’t NVIDIA’s first such joke in recent years, although Ampere mobile units for laptops have bounced back last time.

Over ten days or so ago, we didn’t want to believe the rumors about two versions of the GeForce RTX 4080. Meanwhile, NVIDIA offered a simple trick to shift the price shelves for the next generation, while not having to keep dropping GeForce RTX 3000 graphics cards. .

NVIDIA AD102 – Characteristics of the Ada Lovelace top-end chip. The full core went almost to the professional RTX 6000 card

Not many people (including our editorial staff) were impressed by what was shown yesterday on the occasion of the NVIDIA GeForce RTX 4080 card. The desktop series was split into two completely different systems, not only with a different TDP, but above all with a completely changed setting. The NVIDIA GeForce RTX 4080 16 GB (the only two cards that will be offered in the Founders Edition) is based on the AD103, while the GeForce RTX 4080 12 GB (available only from AiB partners) uses the AD104 graphics chip. For people who use desktop cards, this is at least an odd and debatable solution, but laptop owners faced a similar situation when GeForce RTX 3000 mobile systems debuted. There it was said that power consumption was reduced, but in the case of desktop computers, such a reason would not apply.

NVIDIA Ada Lovelace – Characteristics of 3rd Generation RTX Cores and an Overview of Omniverse RTX Remix Technology

Interestingly, the specifications of the NVIDIA GeForce RTX 4080 12GB card have already appeared several times in the rumors, mostly sourced from kopite7kich. According to his information, it will be a model with a lower number (GeForce RTX 4070). Perhaps at some point, NVIDIA didn’t actually plan to change the name, but the crypto market crash, Ampere chip sale, and Ethereum transformation led the manufacturer to implement a different plan. A game that players do not like, because the new generation price shelves have been raised to ridiculous levels.

NVIDIA DLSS 3 – Only GeForce RTX 4000 card holders will benefit from the new version of AI technology. We know the prices of Polish cards

Looking at this presentation and the specifications of the individual cards, it can be concluded that at the top it was decided to aggressively raise the prices of GeForce RTX 4000 systems, as long as no further price cuts were allowed for Ampere systems. These have already seen significant reductions in recent months, but stock levels are still phenomenal. The introduction of the new generation at the same MSRP prices as the GeForce RTX 3000 series could lead to even larger price cuts for existing cards, likely to deepen the decline in margins. As a reminder, with its latest financial statement, NVIDIA confirmed a very significant drop in gross margin, revenue in the gaming division, and net profit. What we see now is a desire to improve these results. We will likely find out if the plan will work when we analyze the next financial statement.

NVIDIA GeForce RTX 4090 and GeForce RTX 4080 – Presentation of new generation graphics cards. Specifications, price and performance

We expect the prices of current Ada Lovelace models to drop not before the surplus GeForce RTX 3000 cards in stock. Perhaps only then will we hear about the weaker GeForce RTX 4000 models. However, the question will be NVIDIA’s next moves? We received neutrals in the form of a GeForce RTX 4080 12 GB on the AD104 core and 12 GB of memory on the 192-bit (!) Of course, there is a possibility that the future GeForce RTX 4070 will have similar specs (perhaps with the truncated AD104) and with the same memory configuration. In a few months, NVIDIA will have two solutions ahead of it – either withdraw the GeForce RTX 4080 12GB (when the market situation stabilizes and/or enough people buy this model at full price) and replace it with a GeForce RTX 4070 with a similar. Lower specs and price…or a larger GeForce RTX 4070 with 10GB of memory on a 160-bit bus and a more compact AD104 chip. Finally, it’s still the GeForce RTX 4060 – we don’t have much to rely on 12GB of memory. Everything indicates that based on the AD106 chip, this card can in fact receive only a 128-bit bus (and therefore 8 GB of memory). The coming months will show NVIDIA’s next moves, but for now the Ada Lovelace generation has been presented with great distaste, which turned out to be the GeForce RTX 4080 – its variants, differences in specifications and, above all, prices. Surprisingly, the top-of-the-line GeForce RTX 4090 has “only” $100 more than the GeForce RTX 3090. Despite the very high price, of the three, it appears to be the most profitable GPU model today – if we’re aiming for top performance.

Source: PurePC