Nvidia RTX 4070 could be a powerful GPU – but may not arrive for some time

RTX 3070 Ti Close Up Showing Fan and Logo
(Image credit: Nvidia)

Nvidia’s RTX 4070 is in the pipeline – we know that, of course – but it might not be here for a good while yet, and there are doubts over how powerful the graphics card will end up, too. It could be a beast of a mid-range GPU, mind (more on that later).

This is the latest from a popular YouTube hardware leaker, Moore’s Law is Dead (MLID), whose most recent video discusses a whole raft of aspects relating to the RTX 4070. Keep that saltshaker handy as always with material from the GPU grapevine.

One of the main points made here is to not expect this graphics card anytime soon. MLID reckons that a possible launch timeframe is the end of Q1, meaning March, so that’d be five months away still.

The theory is that Nvidia still has a lot of current-gen RTX 3000 stock to shift, so to ensure those Ampere GPUs are sold through, Team Green will purportedly be throttling the amount of RTX 4090 and 4080 boards that hit the shelves as this year progresses (and we enter 2023).

One source told MLID that this problem should be worked out by the end of Q1, hence this being the probable launch date for the RTX 4070, which will only come out when Ampere inventory is sufficiently cleared. Worryingly, end Q1 is the best-case scenario the rumor monger reckons, so we could see an even longer wait for this (relatively) more affordable Lovelace GPU.

The slight caveat here is that a ‘paper launch’ could well come a fair bit earlier – maybe even at the close of 2022 – but that would represent just a reveal and maybe some minimal initial inventory. So, we’d be in that familiar situation where a GPU is launched, sells out instantly – with scalpers involved no doubt – and it’s still a waiting game for the real volume launch of the product.

Another key takeaway from this latest video is that the RTX 4070 still sounds very much up in the air as to the spec. Apparently what won’t happen is what some have suggested elsewhere: that Nvidia might simply take the RTX 4080 12GB and push that out as a 4070.

Mainly because work hadn’t started on actually making RTX 4080 12GB cards, or not all that much had been done anyway, before it was canceled. Remember, this was an add-in board partner product only – meaning Nvidia wasn’t making its own Founders Edition. And all this is partly why Nvidia did cancel this card – while it was relatively easy to do so, following the negative reaction to the GPU as being seen as an RTX 4070 that had been disguised as a lower-tier RTX 4080 in order to put a heftier price tag on it.

And because of that perception, that’s exactly the reason Nvidia can’t simply wheel out an RTX 4070 with the same spec as the 4080 12GB – because then it would be pretty obvious that this was indeed intended to be the 4070 in the first place, as many (including ourselves) previously theorized.

The upshot is that the spec for the 4070 will certainly need to be different to the ‘unlaunched’ 4080, even if it will use the same chip (AD104).

Nvidia RTX 4090 Close-up

(Image credit: Future)

Spec matters

Now, as to exactly what that spec might be, MLID believes Nvidia doesn’t know itself yet, and it’ll depend on how strong RDNA 3 graphics cards turn out (the latest rumor is RX 7000 GPUs could be more powerful than many have thought, by the way).

MLID further theorizes that we could see a weak sauce RTX 4070, which would be a 10GB model with a cut-down amount of CUDA Cores (not the full 7,680, but perhaps 7,168, as that’s a core count which has been floated in the past for the 4070 on the grapevine). This would be the path Nvidia might want to take if RDNA 3 doesn’t look so hot – either from a performance standpoint, or moreover, due to its price/performance ratio or overall value proposition – and indeed if Ampere sales go more swiftly and better than expected.

Another path MLID foresees as a possibility is Nvidia making an RTX 4070 which is a 12GB card and almost the same as the canceled 4080, but cuts down the CUDA Cores a little (maybe only to 7,424 for example). This would allow Team Green to argue this isn’t simply a repositioned 4080 12GB, and carry that perception through, even though we all know what has really gone on behind the scenes, let’s face it. This would be the route if AMD RDNA 3 comes out reasonably strongly upon launch, in theory, and Nvidia feels the need to be more competitive.

Another option MLID mentions is if RDNA 3 outperforms expectations for performance and value (as per that latest rumor). In that case, Nvidia might want an RTX 4070 Ti which actually outdoes the 4080 12GB (maybe with faster VRAM, he suggests).

The leaker says this is the least likely possibility, and indeed, we really can’t see this happening. (“Oh, here’s an RTX 4070 variant which is actually faster than the 4080 lower-tier model we revealed-then-ditched not so long ago” – it makes no sense really. Although you could argue that Nvidia has made no sense with some moves in the past, too; but that’s another story).

Analysis: Patience is a virtue

What to make of all this? Well, the overall gist is that Nvidia is still very much deciding on the positioning of the RTX 4070, and may not make up its mind until after RDNA 3 GPUs have been pushed out. So that necessity is further delaying the RTX 4070, which still needs to give room to RTX 3000 models to sell through, and so we shouldn’t expect the 4070 for quite some time.

In short, much as we might want the RTX 4070 (and indeed 4060) to come out soon, theoretically we’re in for quite a wait – maybe until March 2023, or even further out. A wait that could be made all the more painful because based on some pics MLID also provides for the RTX 4070 – purported images, add extra salt here, naturally – it looks like it sports a large fan and streamlined airflow, with rumors that in testing, the graphics card is staying nicely cool.

We could be looking at quite a GPU, then, if Nvidia decides to take the path of a more powerful RTX 4070 as discussed above – and it’s a well-cooled, relatively high performing graphics card, that’ll have gamers champing at the bit, of course. Except we’ll have to be patient for the GPU to turn up, and indeed perhaps very patient…

Darren is a freelancer writing news and features for TechRadar (and occasionally T3) across a broad range of computing topics including CPUs, GPUs, various other hardware, VPNs, antivirus and more. He has written about tech for the best part of three decades, and writes books in his spare time (his debut novel - 'I Know What You Did Last Supper' - was published by Hachette UK in 2013).