Nvidia Lovelace RTX 4000: everything we know so far

The long-awaited successor to the Nvidia Ampere series of graphics cards has been officially announced, with the next generation being teased for release in mid-to-late 2022. Codenamed ‘Lovelace’ after British mathematician Ada Lovelace (who is considered to be the first computer programmer), you might better recognize these GPUs as the GeForce RTX 40 series.

With the Nvidia hack earlier this year, we had more leaks and rumors than we anticipated, so we had a pretty good idea of what to expect from Nvidia’s Ada architecture, and when to expect it.

But when will that be? It’s unclear when Team Green will unveil its shiny new architecture to the world, but it’s likely that Nvidia will host a dedicated event for the launch rather than piggybacking off an existing conference, as it did for Ampere back in 2020. 

We did hope we might be thrown some scraps during the Nvidia Computex 2022 keynote, though this was more centered around AI and data center innovations rather than PC gaming hardware.

The current estimated release timeline predicts that Lovelace will launch in Q3 2022, sometime between July and September so we don’t have long to wait to get confirmation of many of the rumors floating around online.

Nvidia GeForce RTX 4000 Lovelace: Cut to the chase

• What is it? Nvidia’s next-generation GPU architecture
• When will it be available? Speculated to be Q3 2022
• What will it cost? Unclear, likely similar to the current Ampere series

Nvidia GeForce RTX 4000 Lovelace: release date

The best guess we have right now from the slew of recent leaks and rumors is that the Nvidia RTX 4070, 4080 and 4090 GPUs could arrive as soon as July 2022. This estimate is also backed up by previous generational releases from Nvidia.

The A100 was formally revealed in May 2020, with the consumer Ampere GPUs launching in the form of the RTX 3080 and RTX 3090 about four months later.

If Nvidia follows a similar release schedule with Ada Lovelace GPUs, we can expect the RTX 40-series to arrive sometime between July and September, which would suggest that Nvidia will set its own launch window for Lovelace rather than piggybacking off an event like Computex or CES, but that’s not the only thing we can assume from prior releases.

The RTX 3000 flagship, Nvidia GeForce RTX 3080, was released on September 17, with the Nvidia GeForce RTX 3090 on September 24. The RTX 3070 was released on October 28.

This staggered release is likely to continue with Lovelace, so while the range could grow to feature as many models and variants as Ampere, we’re only anticipated a handful of cards to be available on launch day, with new models released in the following months to plug any gaps in the market.

Nvidia GeForce RTX 4000 Lovelace: price

There’s no officially confirmed pricing for any of the Lovelace GPUs expected to launch the series, though we anticipate that pricing will be consistent with that of the current Ampere range.

The Nvidia GeForce RTX 3080 has an MSRP of $699 (£649, about AU$975), while the Nvidia GeForce RTX 3070 officially comes in at just $499 (£469, about AU$680). The flagship of the family at that time before the RTX 3090 Ti was introduced, the RTX 3090, was priced at $1,499 (£1,399, AU$2,100). 

The chip shortage, unfortunately, ensured many of these GPUs were artificially inflated so many of us were unable to purchase a card at this price unless you were lucky enough to nab a founders edition directly from Nvidia at launch.

If there is sufficient stock available for this launch, inflation and price scalping could be better avoided, so even if the GeForce RTX 4080 will have the same MSRP as its younger counterpart, these GPUs could be more affordable if history doesn’t repeat itself.

There have also been concerns that because this series is rumored to offer twice the power and performance of its predecessor, it could also be double the price. We’ve seen no evidence to suggest this will be the case, but given the last few years, anything could happen.

Nvidia GeForce RTX 4000 Lovelace: specifications

There are currently very few confirmed details about the specifications or the performance of Team Green’s next-gen graphics cards. All that is known about Nvidia Lovelace is that it will use the 5nm production process and the card is capable of Ray Tracing, a feature that has been present on all RTX branded GPUs from Nvidia. 

As previously mentioned, the most prominent speculation floating around on the grapevine is that the Nvidia RTX 4000 GPUs could be twice as fast – and power-hungry – as RTX 3000s, so you might want to look into upgrading your power supply ahead of time to avoid the inevitable shortage when Lovelace is released.

Still, plenty of other rumors have been circulating about power draw and potential performance, so here’s all the information on what we can expect from the GeForce RTX 4000 series.

Going by the leaked information, there are several models under the 4090 name. Most likely the AD102 will be the flagship model, followed by the AD103 as the next most powerful version. The AD104-106 will probably be the mid-range option, and the AD107 will be for the entry-level market.

It seems that the flagship AD102 GPU model will feature 144 SMs in a single die compared to Ampere’s GA102 84 SMs, making this a jump of 71% and one of the largest in a generation. The AD102 could have as much as 18,432 CUDA cores, a staggering increase of 75% over Nvidia’s forthcoming RTX 3090 Ti which features 10,752 CUDA cores.

Established leaker Kopite7Kimi suggests that this graphics card would use the AD102 chip which will power the RTX 4090. Supposedly the 4080 is set to run with 16GB of GDDR6X VRAM.

What’s also spilled here – and very much the hot topic regarding Lovelace, pun fully intended – is more on the power demands of these RTX 4000 GPUs. We already heard recently from Kopite7kimi that the RTX 4090 could sit at a hefty 600W of power draw (and the aforementioned super heavyweight – maybe RTX Titan – spin on AD102 might just be looking at more like 900W, a truly staggering figure).

With this latest leak, Kopite7kimi claims that the RTX 4070 will sport a TGP or power consumption of 300W. As for the RTX 4080, that’ll supposedly have a “similar TGP to GA102”, the latter of which is the chip that’s the engine of the current RTX 3080 and 3090 models (hence the expectation that AD102 would also serve the RTX 4080, which is apparently not the case).

New photos have emerged that apparently show part of the upcoming Nvidia GeForce RTX 4090 Ti – and it looks like it’ll be a monster GPU that people with smaller PCs may struggle to install.

The images were posted on the Chiphell forums (opens in new tab), and spotted by Wccftech (opens in new tab), and apparently show the heatsink and cooler of the Founders Edition, which if real would be Nvidia’s own version of the graphics card.

We also seem to get a good look at the heatsink, which is huge. The RTX 4090 Ti, if it exists, would likely be an extremely powerful – and power-hungry – graphics card, and that means it’ll need a hefty cooling solution so it doesn’t overheat.