The RTX 4070, yet another powerful Ada Lovelace-based GPU, is about to be released. The card’s specifications and performance details were revealed late last year and have remained mostly unchanged. However, the situation does not appear promising for Nvidia’s impending 70-class product, as it may be slower than the previous-generation RTX 3080 10 GB.
Trustworthy informant and industry insider Moore’s Law is Dead described how Nvidia’s Ada Lovelace scaled in terms of theoretical performance and alluded to the pricing and performance characteristics of the impending RTX 4070. According to the YouTuber, this card will be priced similarly to prior GPUs in this generation. Simultaneously, the leaker claims that Nvidia has yet to confirm the details of its add-in board (AIB) partners.
As a result, the highly awaited GPU is likely to hit stores next week. According to some sources, Nvidia is presently planning an April 13 launch, with the corporation remaining tight-lipped regarding the actual date and specifications. As a result, the community has gone crazy with speculation.
The new Nvidia RTX 4070 may be less powerful than anticipated.
The Ada Lovelace lineup has astonished the community with its performance advancements over the previous-generation RTX 30 series lineup in recent months. In reality, all of the cards released thus far can compete with the RTX 3090 Ti and outperform the Ampere flagship by a wide margin. As it stands, the RTX 4070 is likely to outperform the previous-generation model.
However, recent events indicate the opposite. Lesser variations of the new card may be unable to compete with the RTX 3080 10 GB, which is significantly slower than the $800 RTX 4070 Ti.
The 4070’s specifications and projected performance
The RTX 4070 will be based on a scaled-down AD104 GPU and will purportedly have the same amount of CUDA cores as its predecessor, the 3070. The card will include 12 GB of high-speed GDDR6X memory and will be capable of drawing up to 250W in video games.
| Nvidia GeForce RTX 4070 | |
| GPU | AD104-250-A1 |
| CUDA cores | 5,888 |
| Memory size | 12 GB |
| Memory type | 21Gbps GDDR6X |
| TDP | 250W |
These specifications imply a theoretical performance of 29.15 TFLOPs. In comparison, the RTX 3080 10 GB gives a strong 29.77 TFLOPs performance. While theoretical performance isn’t a guarantee of how the cards will perform in video games, it does provide some insight into what to expect.
Moore’s Law is Dead compared the theoretical performance of the cards to their expected performance. This suggests that the 4070 may be unable to keep up with the RTX 3080 10 GB video card. Yet, the 4070 has some advantages. Because to its frame generating mechanism, it supports DLSS 3, which enables significantly better framerate improvements than DLSS 2 on the 3080.
Pricing
The RTX 4070 was initially expected to be more expensive than the RTX 3080. According to industry sources, the forthcoming GPU will cost $749.
Nvidia, on the other hand, appears to have altered its pricing plans, since the company now intends to offer the card at a considerably lower price point. According to Moore’s Law Is Dead, add-in board partners have already been encouraged to deliberately build their cards in such a way that they remain profitable despite a minor price decrease.
Although the leaks haven’t confirmed the price, the YouTuber believes it will be no more than $649. It’s worth mentioning that the 3070 was released for $499 and received high marks for its price-to-performance ratio. Although inflation and a post-pandemic market influence the pricing of the new model, $649 appears to be a bit on the expensive side.
Leave a Reply