With the NVIDIA GeForce RTX 30 Series being an immense success, PC enthusiasts are now looking to the future with the RTX 4000 Series GPU, and we're all keen to know exactly what price we could be looking at.
The RTX 4000- or RTX 40 - Series is the next in line to wow us, and given the power gap between the 20 and 30 Series, we'd expect a similar performance increase for the RTX 40...and a bit of a price spike too.
Interestingly, some earlier reports and rumours suggested that the RTX 4000 series could be skipped altogether, but NVIDIA did eventually confirm its release in a GTC Keynote.
Here are all the rumours, speculation, and predictions that have surrounded the RTX 4000 line of GPUs in the run-up to its release.
RTX 4000 Series Price
Much like the previous lines of NVIDIA GPUs, we expected the company to launch a collection of GPUs at different price points, which is exactly what they've done.
The starting price for the RTX 4090 is $1599, making it $100 more expensive than its predecessor, the RTX 3090. This price tag also means it's Nvidia's most expensive consumer graphics card to date.
Read More: RDNA 3 GPUs - Everything You Need To Know
Meanwhile, the RTX 4080 has a starting price of $1199 for its 16GB card and $899 for its 12GB card, which is once again a significant increase on the RTX 3080.
As for a potential RTX 4070 and 4060, we haven't yet received any official confirmation for them yet, but it's almost certain they will be less than the RTX 4080.
RTX 4000 Series Release Date
It took a while for Nvidia to announce a release date for their RTX 4000 Series GPUs, and initially there were plenty of leaks and rumours.
Some suggested that because there was an excess supply of RTX 3000 series cards Nvidia needed to sell, the launch of the RTX 4000 series would be delayed from September - as was originally expected - until October.
It seems this turned out to have some truth to it, as Nvidia finally confirmed an October launch date.
The RTX 4090 was released on the 12th of October and, whilst there's no official release date for the RTX 4080, the company mentioned during their keynote that it could be expected sometime in November. We're still waiting to hear more on that, though.
As per VideoCardz, leaker @momoom_us has discovered that power supply maker Enermax has eight new graphic cards up for selection, which could relate to the release of either the RTX 4000 Series or the Radeon RX 7000 Series.
If it has anything to do with the RTX 4000 Series, though, it could mean we can expect an RTX 4070 and 4060 soon - neither of which have been officially announced by Nvidia yet.
RTX 4000 Series Specs
When it comes to the specs themselves, again, we're looking at a slight mystery for now, but we can make some informed predictions based on what we know so far...
If a number of leaks are to be believed, we could be looking at the following:
- 12 Graphics Processing Clusters
- 18432 CUDA Cores
- 96MB L2 cache
- Potentially up to 92 Teraflops of computation
- 2.5GHz clock speed
Twitter user Qbitleaks recently posted a leaked image that contains Nvidias test run comparisons of Lovelace's performance against the previous ampere generation, and it seems to suggest a roughly 50% improvement.
In a general sense, the RTX 4080 and 4090, according to leakers, are expected to feature 71% more CUDA cores than their predecessors, that is, if they are based on Nvidia's most powerful AD102 dyes.
ISpeaking of the flagship AD102 chips, the RTX 4080 and 4090, according to leakers, are expected to feature 71% more CUDA cores than their predecessors, that is, if they are based on Nvidia's most powerful AD102 dyes.
Reliable leaker kopite7kimi had noted that the highest-end Nvidia RTX 4000 GPUs based on the AD102 dyes may be set to smash the 100 teraflops barrier to provide competition to AMD's upcoming Radeon RX 7000 series in what he's called an all-out 'gaming war'.
kopite7kimi has also floated the idea that there is allegedly a card in development based on the 'full fat' AD102 dye, complete with 900W TGP and 48GB DDR6 VRAM.
If this is the case, it would make sense for it to be the RTX 4090 Ti, given these specs are doubly powerful than what we believe the top-end RTX 4090 to be, although it's worth noting that, according to the leaker, 'no one knows whether it will become an actual product' due to its double 16 pin power connectors. We can remain hopeful.
Recently. specs for the purpoted RTX 4090 Ti card have leaked, according to kopite7kimi, and it looks to back up their previous leaks concerning a potential 'beast' of a card, although with some revisions. It's expected to come complete with a 800W TDP, alongside 48GB DDR6 VRAM with a speed of 24Gbps.
Intriguingly, kopite7kimi has also noted that the RTX 4000 series of cards may well stick with PCIe Gen 4.0 support, as opposed to going with PCIe Gen 5.0, which may be from the standpoint of practicality.
A recent leak from kopite7kimi has provided a new set of specs for the RTX 4090 that put it as being double the power of a 3090. His Tweet suggests the card will have 450W TDP which is a step down in power consumption from the long-rumoured claims of Nvidia's top-end card pulling 600W or more.
In addition, it should have 24GB of GDDR6X VRAM with a speed of 21Gbps, which is exactly the same as an RTX 3090 Ti, and a shader count of 16,128 represents a large increase over Nvidia's current top-end cards.
Speaking of the RTX 4090, kopite7kimi recently noted that the clock speed of Nvidia's top-end card could be set to get even closer to the elusive 3GHz mark. They've highlighted in a recent specs roundup that its base clock may be 2235 MHz, with a possible boost up to 2750 MHz. For reference, the 3090 Ti has a base clock of 1670MHz with a boost of 1860MHz, respectively.
According to XpeaGPU, an AD102 GPU has been spotted running the 2019 game Control in benchmarks and has achieved 160fps at 4K with Ray Tracing on, which is double that of the RTX 3090's FPS, offering more evidence to the claim that the new AD102 card will be doubly powerful than the existing 3090.
On the point of benchmarks and power, kopite7kimi has noted that both the RTX 4080 and RTX 4070 have been spotted in actual hardware being benchmarked with 3DMark Time Spy. Within the benchmarks, the 4080 netted a purported score of over 15,000, which would put it as being 50% better than its predecessor, while the 4070 managed one over 10,000, which is as powerful as a 3090 and 50% faster than the RTX 3070.
Speaking of cards being more powerful, kopite7kimi has noted that the AD104 SKU, referring to the RTX 4070 will reportedly be able to 'easily match' the RTX 3090 Ti in terms of performance, which would make it seriously capable if true.
With that being said, kopite7kimi has revised his information when it comes to the CUDA core count of the RTX 4080 down to 9,728 from the previously expected 10,248. Everything else within their leaks roundup has remained the same, however.
kopite7kimi has also noted that the RTX 4060 supposedly consumes more power than the RTX 3070, which could potentially lead to a TDP of 250W or more, which is a large jump for any -60 series card, and could make it quite the beefy choice.
On the point of TDP and power usage, kopite7kimi has also noted that the RTX 4080 could nearly pull as much power as an RTX 3090 Ti with a potential TDP of 420W. For reference, the RTX 3090 Ti features a TDP of 450W, whilst the current-gen RTX 3080 pulls 320W.
kopite7kimi has recently updated his predictions for the power consumption figures of the new RTX 4000 series cards. The top-end AD102 dye is expected to feature 800W of power delivery, whilst the AD103 desktop chip is slated to pull 450W, alongside the AD104's 400 watts, and AD106's 260W.
Renowned leaker kopite7kimi has also updated his predictions for the power consumption figures of the RTX 4080 and RTX 4070 in a recent Tweet. He has noted the RTX 4080 is expected to pull 320W, while the RTX 4070 is expected to feature 285W. While this is a fair decrease compared to previously leaked figures, the cards are still in line or above their predecessors by a small margin.
kopite7kimi has also noted the RTX 4090 may well reach the 100 TFLOP barrier in terms of performance with a clock speed of 2.8GHz, which reiterates the potentially serious power of these new RTX 4000 cards.
Speaking of updated predictions, reliable leaker kopite7kimi recently updated his expectations for the specs of the new RTX-4000 series cards which dials back previous leaks, and provides the first full set of specs we'd seen for the RTX 4080 - you can check it out below:
We've also seen a leaked design for the RTX 4090Ti Founders Edition which looks to retain the same dual-axial fan design as previous Founder's Edition cards, as well as a metal frame, thanks to a poster on the Chiphell forum (via VideoCardz). You can check out the alleged design below:
In addition, Moore's Law Is Dead claims to have received renders of the RTX 4090 Ti Founders Edition that are allegedly based on an engineering sample. The card looks to be thicker than its predecessor, which may be to account for its higher TDP. You can check out the alleged render below:
Will RTX 4000 have Anti Mining Drivers?
Aside from the performance differences, we could see NVIDIA implement anti-mining drivers into the RTX 40 Series, as mentioned. These cryptocurrency mining restrictions currently only exist in the RTX 3060, which launched after the other RTX 30 Series GPUs.
Designed to prevent users from mining Ethereum on its GeForce RTX cards, time will tell whether NVIDIA includes this restriction in every upcoming GPU it creates. With stock shortages meaning fewer gamers can get their hands on a new GeForce GPU, NVIDIA may be keen to push miners away from its gaming-focused cards.
However, this doesn't mean NVIDIA wants to lock miners out completely. With the launch of its NVIDIA CMP (crypto mining processor) the company looks to at least want to free up more GeForce RTX 30 Series GPUs for gamers.